On February 11, 1920, a federal census taker came through Rosewood, writing down who lived in which household, along with other information.

This census was taken nearly three years before Rosewood’s destruction, but it still forms the best known inventory of the community not long before its doom. Between the 1920 census and the 1923 violence there certainly were population changes, but survivor testimony shows them to be relatively slight. After 1920 a few residents would continue moving out of the shrinking hamlet, or passed away before 1923. Only three others, in two households, seemed to buck the tide of depopulation and moved in, in each case having temporarily left old roots but then returning. These exceptions, with their special histories—including criminal histories—would feature centrally in the 1923 destruction. They were Sam Carter and Sylvester Carrier (not related despite similar surnames), while with Carrier came his wife Gert.

The census enumerator that day in 1920 was Alfred Dorsett, age 27. He came from a white storekeeping family three miles from Rosewood in Sumner. In the folksy politics of Levy County the Dorsetts were apparently well connected, for in 1917 Alfred’s brother Ed Dorsett, after volunteering with local fanfare for World War I, but failing the physical, bounced back to be named as Precinct 9’s draft registrar. And Alf, three years later, would receive a federal paycheck as census taker.

Thus, for purposes of analyzing Rosewood and its 1923 violence, the challenge is to bring in enough background information from other sources, both courthouse documents and living testimony, to be able to mark off what was and was not Rosewood. Aiding this process is the size of the pool of elderly Rosewood informants located since investigation began in 1982. This is perhaps the largest such group ever assembled to illumnate a rural atrocity from the lynching years.

Also important are the background documents. Though no official documentation on the violence itself has survived, there are records on the details of daily life: property ownership, births. deaths, marriages, and scraps of courtroom proceedings. When this body of information—from both living witnesses and background records—is collated onto the census list, a reliable Rosewood demarcation emerges, and has repeatedly tested out. Not only has there been no evidence-based challenge to the resulting picture of Rosewood, but over the years there has been no publicly known objection to this accounting at all—perhaps because any alternative notions about the community rest on myth and supposition, arising from too little real evidence to frame an argument in public discussion.

My list of the 1923 households in Rosewood, compiled by this method, was approved and used in 1994 by the Florida Legislature’s Special Master verdict that decided the Rosewood claims case—though this in itself is not as conclusive as it might sound, owing to a general tolerance for myth in the highly politicized claims case.

The simple question—Which of the 1920 census households made up Rosewood?—might seem to have been one of the first questions asked in the publicity surrounding the claims case, but politics and passion produced a melodramatic vagueness in the record. The 1992-1994 reparations claim was a nationally headlined cause célèbre, repeatedly promoted by the simple and perhaps half-conscious device of leaving specifics about the martyred community in limbo. The alleged size of both the violence and the violated community were encouraged to balloon indistinctly in popular imagination—almost as if spectacular size was required for an atrocity to be considered.

These dynamics were inadvertently brought to light in 1993 when the State University System of Florida was tasked by the legislature to make a $50,000 study of the Rosewood events, as an aid to claims case deliberations. What resulted, in an atmosphere of fervent crusade in great haste, can be judged by looking closely at the following emblematic passage from the study:

“Elsie Collins Campbell, a white woman of Cedar Key, once lived at Rosewood, and was about three years old at the time of the disturbance. She remembered the village as one of green forests. The view is shared universally by blacks and whites when they describe the community’s dominant features.” (page 20)

This news—that there were forests, and that they were green, as recalled from age three—was the $50,000 study’s only overall description of Rose-wood’s “dominant features,” perhaps marking a milestone in the art of being helpfully vague. The less said about the factual nature of Rosewood, the more public imagination could be pointed toward a mythical paradise, in keeping with the political moment of the 1993 study:

Everyone, “blacks and whites,” agreed that the forests were green, “when they describe the community’s dominant features.” This does not even begin to address the problem that the memory from a three-year-old came from a white describing a black community, or that in fact, if one looks closely, the fudged memory came from before the age of three, since the helpful informant had allegedly moved away before the 1923 violence (if she was ever there at all).

When the green forests were penned, the 1992 Los Angeles riot had recently stunned the nation with more than 50 dead, Florida was still smarting from 1989’s Miami-Overtown riot, and European consulates had placed the state on a don’t-go list because of carjackings and murders of European tourists. Florida was strenuously looking for some happy green forests.

The pattern extended to the study’s treatment of the 1920 census. Rather than attempting to deconflict and explain the census’s undifferentiated list of households, the study simply transcribed the whole mass of raw sheets from Precinct 9 into fully 72 facsimile pages, and affixed them in an enormous, nearly indecipherable appendix to the report—essentially a blind data dump, with no labeling and no hint as to which names and households in the pages did or did not represent the community of Rosewood.

An impression was fostered (and was placed in the report) that the entirety of Precinct 9 had been Rosewood—which would have made Rosewood more than five times the size it really was. This, in turn, would have elevated the 1923 destruction of the community to the epic scale of killing fields worldwide.

The census manipulation eventually grew so obvious that the authors of the study were forced to apologetically retract it in print. Mostly screened from public view by a cheerleading press, there was occurring in Florida not only a groundbreaking experiment in racial financial compensation, but also a demoralizing collapse in historiography. Subsequent journalistic retrospectives on Rosewood would hastily continue to use the census distortion as if it had never been confessed.

The study’s appendix with its 72 pages of 1920 census data was not posted online until 2013, but now might serve as a warning about information from the claims case. In addition to the mystifying lack of labeling, the 72 pages contain various factual difficulties.

The present examination's breakdown of the census, given at some length below, labels the enumerated communities in Precinct 9 by means of verification from independent evidence. Demonstrable errors in the census are noted. The Rosewood households in the list are presented in a two-box summary. The links on each household in the two boxes, leading to longer comments on each, will be enabled as time permits. For now, the reader can find the household-by-household comments by continuing to scroll down. Note that this section is exhaustive, in order to demonstrate the solidity of the community's portrait. If you wish to turn to other search directions, here is the Myths and Evidence Index.

The facsimile census pages posted on the Web from the 1993 study can be accessed here, and then searched by typing in a desired page number.

_____________________

According to the 1920 census, the total number of households in Precinct 9, which included Rosewood, was 151.

Of these, 91 households were apparently in what residents considered to be Sumner, and some 27 were in an area generally considered to be the community of Wylly, northeast of Rosewood, while seven additional households, all of them white, were outriders in the woods.

Despite questions about remote farms and ambiguous locations, the present count posits 26 households in Rosewood in 1920, with four of those being white households, the rest African American. These are the households confirmed by collated survivor testimony, and general agreement on them is close.

Thus, the total number of African American households in Rosewood—the ones that would meet the mob aggression of 1923—formed less than one-fifth of the population of Precinct 9. When the university study for the claims case was presented to the Florida Legislature in December 1993, mistaking all of Precinct 9 as being Rosewood, it caused Rosewood to look more than five times its actual size. For purposes of giving a few long-suffering elderly survivors some compensation (which made scarcely a pinprick in Florida’s budget), this error was no great sin. But it was a less visible attack on an undefended player: the factual accounting of history.

There are reasons why, nonetheless, the present pages use the 1993 transcription from the claims case as a guide to Precinct 9 in the 1920 census. First, the transcription has been usefully posted online by the State Library of Florida, making it available to all.

But second, another kind of value is more unusual. The 1993 transcrip-tion is unfortunately laced with errors, and as these errors are pointed out by the survey here, they provide a lesson showing how official-looking reportage can go awry. The errors stem from hasty interpretation of hand-written census sheets from 1920. The size of this pattern reminds of larger difficulties in the 1993 State University System report, of which the transcription was a part. In haste amid political urgency, that report became an instrument to promote the reparation claim, embracing a number of myths about the Rosewood events and distorting various details. The errors in the census transcription are small compared to some in the report itself.

Below is Rosewood as census enumerator Alfred Dorsett found it in 1920, in households #92 through #116, and then skipping to #144:

Next is the full inventory of 151 households found by Dorsett in all of Precinct 9. The transcription of the Precinct 9 census covers pages 1 through 71 of the appendix in the 1993 academic report:

Page 1 of the transcription begins in Sumner, starting with the household of the census taker himself, young Alfred Dorsett. This household is headed by Alf’s widowed mother Katie Dorsett. All of the six households on this opening page are listed as occupation either “farming” or “none.” Dorsett and his neighbors lived in the freehold portion of Sumner, just east of Cummer Lumber Company property. One of the heads of household on the page, L.P. Smallwood, was at one time Sumner’s postmaster. Page 2 is a continuing bleed of information on these six households, because the long horizontal census sheets would not fit on a single typed page in the 1993 study appendix.

Thus, the next households begin on page 3, with their information continuing across page 4. Of the two households listed here (and the continuation of a third), the second one (household #8) is the home of D. P. “Poly” Wilkerson, the constable and former deputy sheriff who was killed on the night of January 4, 1923, while breaking down the door of a house in Rosewood, thus opening the mass rampage phase of the 1923 violence. Wilkerson’s son James, listed here as age 4, would declare to me in 1994 and testified in the Rosewood claims case.

Then on appendix page 5, with information continuing across onto page 6, there are four households that are still in Sumner, as the census taker veers onto the edge of Cummer Lumber property to record the outlying residence of Cummer supervisory employee R. A. Pope. His daughter Stella, age 2 in this 1920 census, would reminisce in the 1980s under her married name, Stella Herbert.

Next is appendix page 7 (with its information continuing across page 8): The census taker enters the white quarters street of Sumner, though a bit haphazardly. Among the three households here is that of Ollie Surles, assistant engineer at the Cummer sawmill, whose sister Edith Surles, listed as age 13, would in both 1982 and 1983 declare to me about the 1923 violence in Rosewood, whose center was three miles from her home in Sumner.

Then on appendix page 9 (with its information continuing across page 10), comes one of the more confusing census pages. Still in Sumner, and moving to the south end of the white quarters street there, it finds two separate households of Taylors, not related to each other, but one of them would be involved in the Rosewood violence. The first household on this page (#17) is that of mill worker Clint Taylor, with no known relevance to the Rosewood events. But living across the street from him, on the east side of the street through Sumner's company quarters, and at the southernmost end of that street, is census household #18, listed as being headed by Henry Taylor.

“Henry” was the middle name of James Henry Taylor, the millwright whose wife Fannie became a 1923 focal point. The Taylors at the time of the census in 1920 had only one infant son, whose confusing name, “Berness,” apparently led the census taker to label him as “Bernice,” and as a “Daughter.” The 1993 University Team report in which these unlabeled census pages form an appendix scoffed at protestations that Berness was a son, blindly holding the census to be infallible. The name was pronounced “BURN-us,” and by 1923 the Taylors would have two young sons, both of whom I interviewed in adulthood in 1982. Each said he was too young in 1923 to have memories of the Rosewood events.

Page 11 is still in the Sumner quarters street, though skipping some houses. Household #22 here, listed as headed by “Wanda Murphy,” “husband,” should read “Hunter Murphy.” In the excitement of January 1, 1923, Alma Murphy, Hunter’s wife, was said by gossip to have arrived in the beginning stage of the Rosewood violence, allegedly spying what she felt was evidence, as she clapped an empty washtub over a supposed footprint in the Taylors’ yard.

Page 13, still on the Sumner quarters street, finds the census taker doubling back, out of sequence in the houses. Household #24, C. B. Gray, was Sumner’s company- employed doctor. By the time of the 1923 violence, Dr. Gray will be gone, replaced by Dr. A. B. Cannon, the physician who reportedly examined Fannie Taylor after her ordeal on January 1, 1923. Next, household #25, “O. South,” should read “C. C. Smith.” Smith’s wife Elizabeth would be the only known neighbor-eyewitness to the Taylor alarm who survived into the 1980s. I obtained her confusing story in 1982.

Page 15 begins in the Sumner quarters street but by the last of its three households has moved back into the freehold portion of Sumner, outside company-owned land. There are various typing errors here, perhaps coming not from the census but from the 1993 University Team’s transcription. The last name on the page, “Johns, L. L.” should be labeled not “Cousin?” but “Boarder.” Lottie Luconie Johns, born 1880, was Sumner’s justice of the peace. He presided over the January 2, 1923, inquest into the murder of Sam Carter, said by various eyewitnesses to have been perpetrated by mob member Bryant Hudson, though the inquest found “death by unknown hands.” By January 5 or 6, L. L. Johns was reportedly accused by mob members of trying to hide a refugee fleeing from Rosewood.

Page 17, four households, is so poorly transcribed that it might be incomprehensible without background knowledge. Now at the north end of the Sumner quarters street, it finds household “Marthaw, W. L.,” referring to W. L. Markham, manager of Sumner’s company hotel and meat market. Living with Markham is his stepson, age 16, listed as “Phanula, Ernest.” This is long-lived Ernest Parham, who would be a star witness in the 1990s claim case. I also interviewed Parham extensively, as well as his brother Leslie.

At page 18, the appendix page sequencing reverses. Heads of household are now listed on even-numbered pages. The first eight names on page 18 are continuations from the Markham hotel, listing single white men who board there. One of these, “Studstill, Ralph,” is in fact Minor “Cephus” Studstill, age 21 in 1920, who on January 4, 1923, would suffer a flesh wound in the pivotal Rosewood shootout. By the end of page 18, the census taker has veered into Sumner’s other company-owned quarters, a larger but more rustic one for African Americans. The smaller number of whites living on the segregated main quarters street are supervisory personnel or skilled craftsmen. Sumner’s larger number of African Americans are employed only as manual laborers.

Page 20 then presents five households, all in Sumner’s segregated black quarters, with 18 named individuals. A resident of those quarters, Robert Missouri, would survive into the 1980s and was interviewed by me, though Missouri was in Jacksonville during the 1920 census and is not listed by it in Sumner. Sumner’s workforce, indispensable to the running of the Cummer sawmill, were protected from the 1923 violence, which was mostly confined to Rosewood three miles away. Sumner’s African American residents did not witness Rosewood’s destruction.

Page 22, still in Sumner’s black quarters: five households and 18 listed names. Page 24 is also that quarters: five households and 18 names. Then on page 26, continuing in the African American quarters of Sumner (six households, 18 names), notable is household #58, occupied by Sylvester Carrier and his wife Gertrude King Carrier. By 1923, the couple would move back to their original hometown, Rosewood. Released from prison in 1919 on a livestock theft charge (allegedly theft from an African American neighbor), Sylvester Carrier worked for a time as a fireman on a Cummer log train bringing raw material to the sawmill. However, he then lost his job and his company-owned housing in Sumner, reportedly after a quarrel with a boss. The overall supervisor of the log train was logging superintendent Henry Andrews, who was notorious among workers for abusiveness. The quarrel may help explain why Andrews, along with constable Poly Wilkerson, helped lead a group of whites to Carrier’s house in Rosewood on January 4, 1923. After a dog was killed on the porch and a bullet through a window killed Sylvester Carrier’s mother, Andrews and Wilkerson were both shot fatally as they broke through the front door.

Page 28, still in Sumner’s African American quarters: four households on this page, 18 names. All of these residents were in suspense as the Rosewood violence flared nearby, but they were not involved. Pages 30 and 32, more quarters: each page with four households and 18 names. In household #73 is former Rosewood resident Charles Williams, who had run a grocery and moonshine shop in Rosewood before its decline, as the majority of residents, including Williams, moved out. The 1920 census finds him alone; his wife Rita periodically fled heavy physical abuse from her husband. By 1923, however, they would be together again in the quarters, when Rita’s father, James Carrier of Rosewood, would seek shelter there, but would wind up in the hands of mob members, fatally.

Page 34: Sumner’African American quarters continues until household #81—four households, 18 names—then back into the white, non-company area with one household. Page 36 continues with families of old white settlers—the Millers, the Fords, the Ingrams—on the eastern edge of Sumner bordering Rosewood. In household #82 is Oliver Miller, age 2, who in the 1980s and 1990s would share family stories, though he could not remember back to 1923. In Household #83, farmer Jessie Ford had lost a yearling heifer two years earlier to an alleged impulse theft by George Carter, the son of Sam Carter of Rosewood, forming part of the background to Sam Carter’s murder on the night of January 1, 1923. George Carter was said to have chased down the animal and killed it with his bare hands. Accusations at Rosewood in 1923 spoke vaguely of a son that Sam Carter had allegedly helped to escape.

Page 38 continues with more white farm families outlying from Sumner—four households, 18 names. Page 40 ranges to an area called Geiger Creek (though unlabeled in the census), with another page of white farm families: 2 households, 18 names. In household #89 is Bryan Kirkland, listed almost correctly as “Kirkland, Bryant.” This woodland resident, age 23, would be in the group on January 4, 1923, that backed up Poly Wilkerson and Henry Anderson before those two leaders were killed by Sylvester Carrier at his mother Sarah’s house. Bryant Kirkland was wounded in the shooting. In 1983 I extensively interviewed his younger brother Fred kirkland, listed as age 12 in the 1920 census. Fred Kirkland would appear in 1983 on 60 Minutes. But by 1993, when he was found by the “university team” asigned to make a Rosewood study, this informant was bedridden in a nursing home and too disoriented to declare. Nonetheless, the team's 1993 report presented what was in fact only a fruitless attempt to interview Kirkland as if it represented testimony.

NOW APPENDIX PAGE 42 BEGINS THE AREA KNOWN TO RESIDENTS AS ROSEWOOD—THE AREA TARGETED BY THE 1923 VIOLENCE:

Here, land ownership and family ties defined a widely scattered hamlet whose residents self-identified as living “in Rosewood.” The area stretched across two map sections, each a mile square: Sections 29 and 30.

On page 42 of the census appendix, after a three-name continuation of surrounding white residents, Rosewood begins with household #92, listed as being headed by Nathan Robinson, while “Carter C.” is listed strangely below him as “Cook.” In fact, the two-story residence here, dating from an 80-acre federal homestead grant in the nineteenth century, is Cornelia Carter’s home, while Nathan Robinson is her widowed son-in-law. Also living there are Robinson’s married daughter Sebie (listed as “Ceba”) and her husband, as well as Robinson’s youngest daughter, Alanetta “Roby” Robinson, age 4—who would pose a conundrum for the claims case of the 1990s.

Other Rosewood survivors would insist that between 1920 and 1923 the Robinsons moved north to Chiefland, making Alanetta Robinson Mortin ineligible for a $150,000 state reparations payment in the claims case. I was assigned to test this claimant as she insisted she had not moved. The story failed various tests of reliability, but I advised the Attorney General’s Office to give the elderly claimant the benefit of the doubt, since the census said she was there at one point. In later years, as an emerging “last survivor” drew heavy journalistic interest, interviews revealed untrue statements about Rosewood from this claimant, and the implication seemed confirmed by further research, saying that here was another of the unacknowledged ironies in the claims case.

Next on appendix page 42—now that the census taker has entered Rosewood—is household #93, headed by Perry Goins. This household and the 65-acre tract of Goins land it anchors would be at the center of a main Rosewood myth, fostered by the excitement of the claims case, and then inflated to epic size by the 1997 motion picture Rosewood. The myth rests on a distortion of time.

Between about 1895 and 1906, it was true that the Goins family, African Americans, had established a thriving business presence in Rosewood, with dozens of employees harvesting pine resin in surrounding forests, for distillation into products including turpentine, as well as pitch and tar for sealing wooden ships.

But the myth used by the claims case transported this bygone face of Rosewood into 1923, claiming that the 1923 violence had attacked not a scattered remnant community but a thriving business center. The 1923 attack was presented as an alleged proof of how African American wealth was not allowed to be transmitted from generation to generation. In truth, the Goins business founder, Martin Goins, died in 1903 and the business then withered, going bankrupt by 1916. Some of the Goins heirs then moved away, but a nephew of Martin Goins, Perry, remained on the 65-acre tract that had once housed M. Goins & Bros. Naval Stores. As the 1920 census shows, Perry Goins continued to live at the old Goins house site with his wife Hattie and their children, but for employment he walked two miles down the railroad each day to stack lumber for the sawmill at Sumner. In 1982 I interviewed his daughter Beatrice, age eight in 1923.

Next, still on appendix page 42, is another household reflecting Rosewood’s decline since the earliest years of the twentieth century (a time when a cedar sawmill had also bolstered the community, until that mill closed in 1911). Originally, the Goins family had begun setting up Rosewood operations in the late 1890s, arriving from their former company compound on 4,447 acres in Hoke County, North Carolina, where the pines had become tapped out. With the Goins migration had come other families sharing a common trait with them: triple ethnicity derived from African, Native American and European roots, so that some neighbors in frontier Florida had puzzled and said the newcomers looked white, though Martin Goins emphasized that he identified himself as a Negro. Census takers labeled some of these families with neither “B” for black nor “W” for white, but “Mu” for Mulatto. No racial hostilities were recorded or remembered in their area, the western half of the two square miles in the scattered Rosewood hamlet. This became the lighter-skinned half of the community (known to some darker Rosewood neighbors as “Yallertown,” for “high yellow“ or light-complected).

One of the families linked to the Goinses was the Haywards, who in those early years around the turn of the century bought an 80-acre tract between the 65-acre Goins plot and the 80 acres of Cornelia Carter (who also “looked white” or “looked like an Indian”). Greeting the census taker in 1920, the old Hayward home (household #94) could still be described in terms sounding superficially like prosperity—not only a house of two stories but downstairs in the parlor was an upright piano. Yet the surviving heir, Mary Jane Hayward, had fallen on hard times, remarrying and fading from community life (the transcribed census had her husband as “Screw, F. W.” but he was F.W. Screen.).

Moving to page 44: The next Rosewood household, #95, is said in the University Team appendix transcription to be headed by “Mouma, John,” but this, less exotically, should be “John Monroe.” When one of Martin Goins’s daughters, college-educated Sophie, had married Monroe, they moved into a comfortable home on a ten-acre plot carved out of the edge of the Goins tract. The census found John and their four children there, but Sophie had died in childbirth in 1917. More than a mile away in the darker-skinned half of Rosewood, Virginia Bradley, Nancy Bradley, and Daisy Mitchell had all died in much the same way, and in the same short period, despite efforts by Rosewood midwife Mary Ella McCoy. The struggling remnant community was a dangerous place to be a mother, quite apart from the racial strife of 1923.

Still on page 44, the next censused household in Rosewood, #96, is listed correctly as Annie Blocker, though neighbors knew this energetic, helpful woman with four children as “Sis” Blocker. She was Mary Jane Hayward/Screen’s sister. Her prospects had soured with marriage to a perennial fugitive from the law, John Blocker, who in earlier years had run a small moonshine shop and card room in Rosewood. Blocker’s place was infamous for an incident, years before, when a traveling gambler had shot neighbor Arthur Bradley in the hand.

All that was gone now, even by 1920 and the census, and by 1923 the shrinkage would be worse. Ironically, the communal decline had cleaned Rosewood up. On both sides of the remnant community timber operations continued to thrive—in Sumner three miles southwest and in a still-bustling turpentine camp called Wylly only a mile to the northeast. Both those places had booming Saturday night “jooks,” makeshift dancehalls giving cramped, violence-plagued solace to the labor force.

But Rosewood, out of sight of both the remaining hotspots on the railroad that served them all, had become a quieter place, notable for households of grandparents doing their best to care for orphaned or abandoned grandchildren.

Soon after the 1920 census, Sis Blocker, too, would become part of the outward migration, leaving Rosewood and moving her children into a shanty in the Wylly turpentine camp. She would be there when the 1923 violence struck Rosewood. She was spared because the violence did not reach into Wylly. Some refugees from Rosewood’s shootout of January 4, would flee to Wylly and hid in a pine grove there. Annie Blocker slipped through the woods to them and brought much-needed food. Ultimately her strong character would win out when she left turbulent John Blocker for good and moved her family far south to the town of Sarasota, where her son Ernest grew up to be a prosperous businessman. Aged 13 in the 1920 census, Ernest Blocker was interviewed by me in 1982 and again during the 1990s claims case. I also interviewed his articulate sister Arthina Bennett (whom the 1920 census had seemed to identify as a boy).

Though the census record doesn’t show it, enumerator Alf Dorsett would make a long leap after the Blocker house. More than three quarters of a mile intervened between the Hayward property, on the western edge of Rosewood, and the site of the next census entry, household #97, which was in Rosewood’s very heart. Dorsett would loop back later to catch houses in between, making the resulting census picture of Rosewood quite a maze. Household #97, however, would one day become the most famous still-extant Rosewood landmark, for that house was not destroyed by the 1923 violence. It would survive, still well-maintained beside a sleepy two-lane highway, into a new millennium.

This was the rambling Victorian home of the Rosewood community’s white storekeeper, John Wright, and his wife Mary (the census, economically, called them both “Wright, M. J.”).

At least three influential whites would be involved in hiding Rosewood’s 1923 refugees from the mob rampage of January 4-7, 1923. These rescuers included Sumner’s mill boss Walter Pillsbury, turpentine camp owner D. P. McKenzie on the other side of Rosewood in Wylly, and, most instrumentally, John Wright in Rosewood itself, the community’s lone remaining storekeeper.

Mythology about Rosewood has often presented it as having two stores as it met 1923, the white-owned Wright store and a black-owned counterpart belonging to resident Bacchus Hall. In truth there was only one remaining store; Hall’s store had reflected the dying timber economy and closed down well before 1923, as his children Sam, Wilson, and Margie would confirm in the 1980s and 1990s. Bacchus Hall himself died in 1919 and his empty store building caught a spark and burned down, its flames unrelated to the later racial conflagration of 1923.

Myths in this regard portray the Rosewood that met 1923 as an economically thriving place of black-owned businesses, supposedly showing that wealth and influence could soon have been accumulated here—if not for the artificial intervention of systemic racism in the form of mob attack, which allegedly cut short a promising future. This reverses the real history: Dying Rosewood arguably was a tempting target for mob aggression precisely because it already looked half-abandoned. There were no African American-owned business at all. If, by contrast. the myths were to say that even in spite of Rosewood’s difficulties it certainly did not deserve to be attacked, and that it remained a place of hard-working people with close family ties, this would be closer to the real history.

And now, finally leaving behind appendix/census page 44, we move deeper into the heart of Rosewood on page 46. The census taker now loops to two of the few other white households in the Rosewood area. Household #98, Luther/“Arthur” Ingram, was so far back in the woods that it was not usually remembered as being part of the community. But #99 was near the community’s center, a quarter mile or so back down the railroad from John Wright’s store. This was the home of white World War I veteran Rob Ingram. Close by lived a third Ingram, Rob’s brother Will, but Will would not be listed by the census until far over on transcription page 50, as household #106, when the trudging census taker doubled back to pick up loose ends. The resulting hopscotch pattern does not make a very easy road map to a cryptic atrocity, and this was compounded in 1993 by errors and mistypings in the Florida “university yeam’s” transcription.

Relations between the white Ingrams and their more numerous African American neighbors seem to have been generally peaceful, barring the kinds of arguments seen all over the woods involving suspicious disappearances of free-roaming hogs. Indeed, suspicions of that kind were so common in the old frontier South that they had kicked off West Virginia/Kentucky’s Hatfield and McCoy feud in 1878. In Rosewood a shoe would finally drop in 1937, when Will Ingram, having been undisturbed by the 1923 violence, would be arrested for stealing a hog. Allegations that a son from one of the Ingram households had joined the 1923 mob remained at the level of unsubstantiated gossip. Despite frequent post-millennial media claims that only one house (storekeeper John Wright’s) was left standing in Rosewood after the 1923 destruction, the pattern was not so neat and other buildings survived: the two white-owned Ingram residences, the home of George and Maggie Bradley (rented from a white), Rosewood’s small shed railroad depot, Wright’s store, and two homes owned by African Americans, spared out of unpredictable sympathy or chance: the Ransom and Julia Edwards household and that of Jim and Luvenia Hall. Nor is this necessarily a full count. There is no reliable inventory of the 1923 destruction.

Disconcertingly, page 44 of the census transcription now jogs again, leaving the Ingrams, who lived south of the railroad, to go north of the tracks and push back into the woods, finding household #100, the residence of Ed and Eliza Bradley. We are now in what might be called the dark-skinned half of Rosewood (as opposed to the light-skinned half to the west, linked to the Goins legacy). Though the Goinses and related families were latecomers to Rosewood, arriving in the 1890s, the darker families like the Bradleys were old settlers, their forebears arriving before the Civil War, perhaps as slave labor building the railroad here, its last track laid in these isolated swamps in March of 1861, at the war’s very beginning.

In census records, African Americans were not identified by name here until after the war brought emancipation. Before that they were glossed in anonymous head counts for each slave owner. The first post-war census came in 1870, when emancipated surnames seemed to pop up out of nowhere, one being Barnabas Bradley, designating a child in an undifferentiated census tract looking roughly like the precursor of Rosewood. Barnabas would grow up to be Ed Bradley’s father—and the father of two other main householders of 1923: John Wesley Bradley and George Bradley. A nineteenth-century neighbor of Barnabas, blacksmith Isaac Carrier, similarly started Rosewood’s Carrier line, which married into the Bradleys to produce a tight-knit Bradley-Carrier faction in the old-settler side of the community. This faction was counterpoised, after the 1890s, by the lighter-skinned Goins domain just west in Section 30.

The Bradleys and Carriers of Section 29, around Wright’s store and the railroad’s shed depot, had smaller individual land holdings than in the Goins area (for example, the eight acres recorded as owned by Ed and Eliza Bradley), but as early as 1886 Barnabas Bradley had been a Rosewood church trustee, along with another dark-skinned old settler, long-lived Bacchus Hall, as they founded Rosewood’s Pleasant Field Methodist Episcopal Church.

This wood-frame chapel and a later competitor, the First African Methodist Episcopal Church of Rosewood, would by 1912 be experfiencing painful quarrels and feuding, amid tense awareness that the small timber boomtown was abruptly dying, its over-harvested natural resources exhausted, with the only choice left being to stay and find some way not to starve, or get out. Most white residents of Rosewood, who before that had been a majority, did get out, many African Americans as well. The hard-pressed core that remained were mostly older African Americans caring for bereft grandchildren or adoptees, with fewer and fewer young adults of job-seeking age still on hand.

There was the consolation of a warm family support network, while garden plots together with hunting in the woods could fill some gaps. A few residents made arduous six-mile-a-day round trips on foot to get paychecks in the still-booming sawmill town of Sumner, or tried lower-paying turpentine work in Wylly in the other direction. In 1920, Section 29’s remnant group, wreathed in moss-hung shadows that were dotted with migration-abandoned homes, may have been confusing to an earnest white census taker from regimented Sumner.

Ed Bradley, for instance, was jotted down in the page 44 census list as being aged 20. In fact, voluminous documentary and testimonial evidence leaves no doubt that Ed Bradley was an aging lumberjack who had worked the swamps for Cummer Lumber, though constant immersion had brought “water poisoning” to his feet, an ailment known to later soldiers in Vietnam as jungle rot. At home, Ed Bradley dressed deer carcasses he shot for sale passing train conductors, planted corn and peanuts to fatten hogs penned at the back of his lot (where the smell wouldn’t carry), cut some sugar cane, and had a cow for milk and butter.

Bradley’s industrious wife Eliza had once been renowned in Rosewood for bringing home whole wagonloads of dirty laundry from Sumner to wash for white supervisory personnel there. But Eliza’s pace was slowing and by 1923 the wagon trips had apparently stopped. For years the aging couple cared for two adoptees, orphaned niece Thelma Evans and dislocated Bertha Carrier (listed in the census as aged 19 and 17). One of them, Thelma Evans Hawkins, I interviewed in 1982. Shortly after the 1920 census was taken, Thelma became one more working-age resident who fled the lack of opportunities in the enclave in the swamps.

Though she loved her adoptive aunt and uncle Ed and Eliza, she reluctantly described her role in the household as that of a child slave, not allowed to go to school but kept working at fireplace-heated flatirons to press clothes that her aunt had washed on consignment. Thelma spoke to me with crisp diction and careful grammatical phrasing, but this bespoke a hard journey.

Breaking away from Rosewood as she neared age twenty, she had made a lonely trip 120 miles up the railroad to the city of Jacksonville, winding up as a domestic working for a wealthy white couple, who encouraged her ambitions. She described how tired she felt doing double duty, working all day in Jacksonville and then going to night school, determined to make up for the unschooled vacuum left from her childhood. Late at night, she recalled, she would nod off as a streetcar took her home, and lack of sleep took its toll. Then one day in 1923 she happened to glance at a newspaper rack. The horrifying headline there cried, “Rosewood.” Thelma was beside herself with worry about her relatives as she read about the killing and flames. She knew that the place burning in the headlines was not a pristine paradise of dynamic promise, as a later age would seem to keep imagining, but she also knew it was not a brooding outlaw’s roost, as the 1923 Jim Crow news stories seemed to sneeringly imply.

In the 1940s and 1950s, Thelma Hawkins would help care for (and imparted her crisp diction to) a young cousin fleeing not mobs but traumatic turmoil at home. In 1993 that cousin, troubled Rosewood descendant Arnett Doctor, by then age fifty, would become the “point man,” or chief publicity agent, for Florida’s Rosewood claims case. In a sentimental new era he would tell eager news media that the lost paradise called Rosewood had been a fantastically thriving “Mecca,” drawing opportunity-seekers, he said, from all over the South, “like Atlanta is today.” He further said that at the moment of its destruction this lost Eden had been a place of “mansions...with manicured lawns.” Even years after Doctor’s 2015 death, new bursts of media excitement would continue to preserve some of those fantasies as if they were fact. Such, apparently, is the power of racial anxiety in a nervous nation.

And still, page 46 of the appendix census transcription has another household to explore: #101, the Halls. This dour old nineteenth-century saltbox of a house, a two-story dwelling but with no front porch, projected a desolate appearance. The nine occupants here, long-suffering Mary Ann Hall and her eight children, had often known hunger, as they lived on in the family home left by the late Charles Bacchus Hall, the onetime Rosewood storekeeper. A stern man with white side whiskers, Hall would be remembered for chasing away children who sought to play with his own. Just across a field was the previous census listing, that of Ed and Eliza Bradley, and between the two homes was Rosewood’s one-room school. This is poignant, for in both homes the children were forbidden to go to school, and were made to earn their keep.

In 1983 I repeatedly interviewed Sam Hall, Bacchus Hall’s second son, who was 16 at the time of the Rosewood violence. Sam Hall was a forthright, careful informant who in 1983 I brought, as contractor for 60 Minutes, from his home in Georgia to on-location taping at the site of Rosewood. I also interviewed Sam’s surviving siblings, Doshia, Margie, Wilson and Mary, but Doshia had become senile and Mary seemed to have blocked out many memories, while the remaining two had been small in 1923, with few impressions retained.

The hazing or blocking of pre-1923 memories in ths household was backlit by some troubling documentation, including prior Rosewood censuses in 1900 and 1910. Long-lived Bacchus Hall had married far back in the past, on March 11, 1877, though not to his wife of later years. After his initial marriage to Margaret Dawkins, later censuses showed, Bacchus Hall had begun growing old with her, while into their childless household by 1900 a young ward was adopted: Mary Ann Davis, from an impoverished family elsewhere in Rosewood. By 1920 the seven children then born to Mary Ann knew as a matter of course that she was their mother, but seemed to think she had always been Bacchus Hall’s wife. An explosive scandal had developed in Rosewood after 1900 when the children began to be born, with neglected wife Maggie still somehow living in the same household, according to the 1910 census. The Halls were apparently scornned by devout Methodist neighbors, but seemingly the younger children were never told the details. After 1910, Margaret Hall disappeared from the records and Bacchus, old enough to be Mary Ann Davis’s father if not grandfather, then married her formally. By that time the community was dying, the old entrepreneur was selling off parcels of land, and hard-pressed neighbors were defaulting on credit tabs at Hall'S store, closing it down. Mary Ann Hall found herself caring not only for a house full of hungry children but the dying old man in a protracted illness. Arduous relief came from labor shortages caused by World War I, as Mary Ann Hall went to the Sumner sawmill and “stacked lumber like a man,” her children recalled. This was the reality behind a later era’s media myths saying that 1923 rioters had cut short the bright promise of an allegedly still-thriving black-owned store—which the myths also multiplied into an image of various such imagined businesses, in a “thriving town.”

When the 1923 violence struck, Mary Ann Hall and the six children who were still at home fled into the woods, with rescue two nights later organized by influential whites seeking to outwit white rioters. The rescue arrangement took Mary Ann and her five youngest children to Gainesville 45 miles up the railroad, where they found housing and at last a chance to go to school. I was shocked to hear Margie Hall say that she was not traumatized by their flight, and instead she was “glad to get away from down there...There wasn’t nothing going on.” The youngest Hall daughter, Mary, was described as going through high school in Gainesville.

In 1994 when the Rosewood claims case paid its ground-breaking reparations, a total of nine 1923 survivors were still living, and received $150,000 each. It happened that three of those survivors were all from census household #101, the Halls (though Sam Hall, the most assiduous in reporting the history, had died in the 1980s and got nothing). Thus a third of $1.35 million in Rosewood survivor payments went to a single family, whose balance sheet of long-term losses and gains was disconcertingly mixed.

But now, at last finished with page 46 of the census transcription, we move to the next page of names, page 48, still in the heart of Rosewood. Here the previously seen pattern of a white census taker's misunderstandings reaches a crescendo, complicated further by 1993 errors in transcription. Household #102, “John Hall,” should be James Hall. Jim Hall, an elderly woodsman, was Bacchus Hall’s less prosperous brother, living in a small home nearby with his wife Luvenia (“Rubina” in the transcription). Then after that comes one of the transcription’s more ambitious errors: “Columbus, John” is said to live with his wife “Columbus, Ona,” in household #103. Columbus? No one from Rosewood ever mentioned a neighbor named Columbus. In fact, the weary census taker has reached the free-thinking household of veteran Rosewood hunter and trapper John Coleman, and his wife Emma Carrier Coleman.

In better days, crusty John Coleman had been town marshal of a budding settlement that hoped to grow larger. He was known as an apostate who scoffed at Rosewood’s more numerous church folk, and stayed home on Sundays to shoot pennies in crack-a-loop games with younger cronies. The Coleman home was another whose parlor held a piano, and in this case no dilapidation seemed to apply. In Rosewood's eariier days of roaring saws and steaming turpentine stills, this home had thrown fondly remembered parties, and Coleman could afford the musical accompaniment. Back then, before the surrounding woods were hunted clean, stacks of pelts shipped north from the Rosewood depot could make a man’s living, as the new Chicago catalog houses like Montgomery Ward and Sears, Roebuck bought furs and hides and stitched their own merchandise in the form of fur coats, hats, and many other items. But then Rosewood's wild game had thinned like the clear-cut trees, and old John had more time on his hands for playing crack-a-loop with the boys. His wife Emma continued to make two-mile trips on a back trail west to reach the Goins side of the community, where John and Emma Coleman’s daughter Hattie had married the last Goins holdout, Perry, joining together the darker and lighter communal factions. Over there, Goins children rejoiced to see Grandma Emma toddling down the path, cradling a batch of freshly baked cookies.

But questions about this page of census transcription go deeper than what appears in plain view. Pieces are missing. Hattie Coleman Goins also had a sister, Virginia, who had married less fortunately, becoming Virginia Smith. The marriage to Allen Smith had left Virginia abandoned with three small children, as her husband Allen moved to where the work was, at Otter Creek twelve miles east, and apparently did not look back. Persuasive testimony from Virginia Smith’s daughter, Nettie Smith Joyner, together with other evidence, places that family of four in a small, smoky hovel not far from the larger residence of Virginia’s parents John and Emma Coleman. It may have been an easy house for a hurried census taker to miss, or perhaps they did not move there until after the 1920 census. Its placement, however, was established by testimony in the 1980s before there was a claims case or money incentive to encourage exaggerations. The four Smiths—Virginia, Nettie, Johnnie, and Gilbert—were present in 1923 and fled the mob violence. At one point Virginia reportedly panicked as a white man on horseback galloped toward her, and she left her terrified daughter Nettie standing in the road. The horseman turned out to be a rescuer, apparently turpentine woods rider Jack Cason, and the Smiths seem to have been reunited at Cason’s farm.

Next (passing the invisible gap in the census where Virginia Smith’s one-room household fails to appear) there is a larger and more impressive house, but also warped in transcription. According to the census, household #104 is headed by a fifty-year-old widow, “Edward, Julie.” But in reality, Julia Edwards (her real name) was not widowed and her husband Ransom Edwards was still the household head. Ransom Edwards had seen some disillusioning things in his long life, and perhaps he didn’t care to greet a census taker.

This comfortable farm, with its fruit trees and livestock, had been a hub of bygone hopes in Rosewood, for back in the boom days Ransom Edwards was not only a farmer and woodsman but superintendent of Rosewood’s segregated black school. He signed teacher paychecks in a flowing hand, but also had a close view of the daunting neighbor-against-neighbor feuds which, residents said, tore at the community’s sinews.

No one managed to catch the arsonist who hit the school, though it was assumed to be an internal affair, involving no whites. This was frontier country, and tempers could flare; forty miles north in the community of Newberry a white school was burned much the same way, apparently by a white neighbor with a grudge. Also internal, apparently, and not racially motivated, was Rosewood’s arson of August 19, 1911, burning down the African Methodist Episcopal Church (founded in the 1880s like the Methodist Episcopal church to the east).

At first, doughty trapper John Coleman was accused of the 1911 church burning; notoriously impious, Coleman also had a son who, bullying one of the Carriers, had gone too far and was shot to death. But whatever his history, more credible suspicions soon fell on former church treasurer Richard Williams, who had reportedly been caught stealing from the collection plate. A jury refused to convict.

The pattern of internal quarrels continued, fueling a reputation for violence as the community shrank.

In 1983 I interviewed Ransom Edwards’s grandson Willie Evans, who had lived with Ransom and Julia in Rosewood. He said that his grandfather had stopped going to church, apparently disgusted by the neighborly acrimony. The old man was devout and had a prized collection of religious books, but he seemed to have closed a door. The destroyed church was rebuilt as the First A.M.E. Church of Rosewood, but then its preacher was shot and killed at the pulpit, by one of the Goins brothers, not by an outsider.

On page 48 the last household, which continues into the next page of listings, page 50, is household #105, labeled as “Bradley, West.“ This was John Wesley Bradley, known to neighbors by his middle name, Wesley. He too, like his older brother Ed Bradley in household #100, was a lumberjack for Cummer Lumber, sawing down towering cypress trees to feed the sawmill in Sumner. Each Monday, Wesley Bradley would walk to Sumner and catch the Cummer log train to travel twelve more miles north to a logging camp at Fowler’s Bluff, in swamps along the Suwannee River. Wesley Bradley would stay there felling trees through the week, then returned on Saturdays. His wife Virginia would by then be preparing Sunday dinner in advance, for as John Wesley Bradley’s name implied, this was a household of hard-shell Methodists, who forbade any kind of Sunday work, or even the running and playing of children on Sundays.

The piety did not imply dour gloom, however. Wesley was a hearty jokester, enlivening his crowded Sunday dinner table with expansive talk. Besides housing nine children, this rambling, one-story residence, edged with an L-shaped porch, was said to keep a room available for visiting preachers, whose appetites for free food on Sundays was said to be noted drolly by the householder. One day Wesley was said to call out across the table (when no visiting ecclesiastics were present) to skewer his busily eating son Donarion, proclaiming: “Son, you don’t just eat like a preacher.You eat like a >bishop!” The roar of laughter at the table promptly renamed the diner as “Bishop” Bradley. He seemed to be called by no other first name ever since. Even the 1920 census called him “Bradley, Bishop.”

A voice from this household, that of Lee Ruth Bradley Davis, belonged to one of the first survivors I found as I began investigating Rosewood in 1982. Also tremendously helpful was Leroy “Lee” Carrier, a boarder who at the time of the violence was living in the same household, that of his uncle Wesley. The differences in the observations of such witnesses help to show why a large pool of informants is important for faithful accounting of such a case.

Between the time of the 1920 census and the January 1923 mob attack, another kind of tragedy would strike Wesley Bradley’s household, for Virginia Carrier Bradley, his tireless wife, died in yet another childbirth. Oldest daughter Callie married and moved away, as did oldest son Hoyt, joining the stream of young people leaving the settlement. Wesley’s sister-in-law Maggie Bradley tried to look in on the younger children. The house itself was in limbo, apparently not owned by Wesley but originally church property. Meanwhile, the Cummer operation in Sumner was gearing down for scheduled shutdown in 1926, when its operations, including logging, would move two counties away. Though strained by many pressures, this household had apparently remained apart from Rosewood’s internal quarrels and feuds. In 1923 it would be a target nonetheless.

On page 50 of the census transcription, after the long list formed by Wesley Bradley’s large family, the census taker doubles back to record one of two white households just south of the railroad, as mentioned earlier on households #98 and #99. Geographically, household #99, Rob Ingram, was not far from the household we meet now, #106, that of Will Ingram, Rob’s brother. As previously said, it was Will who would be arrested in 1937 for stealing a free-roaming hog (one of the signature crimes of old Florida), but Rob, also, was well remembered for his woodland “hog claim,” which brought accusations from African American neighbor Sam King, just below.

Before examining that, however, the next entry on transcription page 50, household #107, was the site of a shocking fatality in the 1923 violence, on the night of January 4-5. This is the home of Lexie Gordon. The week-long Rosewood explosion, January 1-7, apparently produced two main arson sprees, the first burning only a few homes on Thursday night, January 4-5, then the second, on Sunday, January 7, making a more extensive effort to burn down the whole community. It was Lexie Gordon’s misfortune to be in the path of the first spree.

As she fled her burning residence she was shot down. This soon led to legends that perhaps dozens of people in Rosewood were killed this way, by “burn-and-shoot” tactics, a mob method made nationally infamous in the East St. Louis race riot of 1917. At Rosewood, however, both survivor testimony and documents leave no doubt that almost all residents were able to flee into hiding very quickly at the first sounds of gunfire Thursday night, so that when houses were burned no one was inside to be shot while trying to get out—except for Lexie Gordon.

Age 55, a widow, she had been sick in bed with “typhoid-malaria fever” (a diagnosis emerging in the Civil War, meaning chronic underlying malaria infection overlaid with an acute case of typhoid). Flames from unidentified arsonists evidently forced her outdoors, where she met gunfire.

This racial killing was ironic because some African American neighbors viewed Lexie Gordon as being white. Very light-complexioned, she had long red hair. but was listed as “Mu,” for Mulatto, in the census. She, too, was one of the 1890s migrants coming with the Goins pine-resin distillery business. The five acres she and her late husband had bought in 1910 lay on the farthest eastern edge of the Goins area, entering Section 29 and the half of Rosewood where older settlers were darker of skin.

A plot adjoining the Gordons’ was bought on the same date by a lumberjack and farmer named Sam King, who had a similar background. The home he built there for his wife Ellen and adopted daughter Eloise would appear in the 1920 census as household #108. In 1982 I interviewed Eloise King Davis (“King, E....Daughter” in the census) and she described in detail a home with a pedal sewing machine for her mother’s seamstress work, a picture of Jesus over the fireplace and a small toy stove for a pampered only child. All went up in smoke as the Kings barely escaped the arsonists of Thursday night.

The 1920 census then found two households, #109 and #110, whose four occupants were said by other residents to have either died or left the area before 1923. But next came a more notable household, #111, headed by the third of the Bradley brothers. George Bradley and his wife Maggie were “the bold Bradleys,” in part because of physical size. George and several of his children, including girls, were said to be over six feet tall. Their home gave them breathing space, for the eleven people listed by the census in this household, including children and grandchildren, were living in what had once been Ford’s Hotel, another reminder of Rosewood’s decline. Near the defunct hotel were vine-covered stanchions left from the old community hub, a vanished cedar sawmill that had given Rosewood its name, before the mill closed in 1911. Big George Bradley continued to be renowned as a hunting guide. At one time wealthy northern sportsmen had come in by train, seeking wild game in the ten-thousand-acre swamp forest called the Great Gulf Hammock, which lapped up to the back door of Ford’s Hotel. There the shadows of the hammock met ever so slightly higher, drier pine forest, forming an interface that defined the positioning of Rosewood.

The census taker was now making nearly his last crisscross through the settlment's bayhead thickets and palmetto fans. The old hotel that was George Bradley’s house was on the south side of the railroad, and census enumerator Dorsett would now follow those rails toward the farthest eastern edge of the community (more exactly, the rails ran southwest to northeast, and Dorsett was moving northeast through eastern Rosewood, toward Wylly). In this direction there were only five more Rosewood homes to visit—but the count has a major gap. The census would list only four of those homes. Moreover, the one missing from the 1920 census—for reasons probably mundane—would be the single most important nexus in the 1923 violence. This dwelling, which for a frenetic moment in 1923 would make newspaper headlines from New York to Los Angeles, was the home of Sarah Carrier.

But first, before examination of that omitted flashpoint, here are the four homes in that direction that the census did list:

Page 54, household #112. Though Sam Jones is listed in 1920 as heading this three-person household, he was gone by the time of the violence in 1923, reportedly having died. His widow Laura Jones lived on in modest circumstances. If a neighboring hunter killed an opossum, an animal disdained as food by many for its habit of eating carrion, tte carcass might be donated to Miz Laura, since she was not so picky. By 1923 Laura Jones’s son Raymond would be old enough to stack lumber in the Sumner mill, walking to work each day with Sam Hall, 15, from the Bacchus Hall household. The long planks on the fifteen-feet-high lumber stacks in Sumner were heavy and dangerous, and Hall would recall an injury as he fumbled to learn the new job, when he "worked all day with blood in my shoe.” Laura Jones owned her eleven-acre house site, paying its yearly taxes.

Page 54, Household #113: This was a faded but neatly maintained two-story residence with a picket fence cloaked by a rambling rosebush. Chickens and children roamed the sandy yard. Facing the railroad on the north side, this residence looked across the tracks toward a mass of trees in the middle distance, the edge of the Gulf Hammock swamp, where children from the house were forbidden to play.

Visible just north behind the house, a rock road would be under construction by 1923, eventually to cross a sodden barrier called Devil’s Hammock and then reach the town of Gainesville (the road that later became Florida Highway 24). Children playing in the back yard of the house occasionally saw convicts in stripes being transported to work sites father east on the road. Florida’s notorious practice of leasing convicts to private employers was coming to an end by the early 1920s, and the state road gangs of later notoriety were arising.

The household here, #113, is addressed in some detail by the present discussion because at least six of its occupants would be caught in Rosewood's pivotal gun battle of January 4, 1923. Two would receive wounds. But that was nearly three years in the future as the 1920 census taker came to call.

The head of this household is listed correctly by the census as James Carrier, though by 1923 James Carrier had suffered two cerebral strokes and was partially paralyzed; household management had fallen to James's tall, rigorously disciplined wife Emma. Both spouses were in their fifties, caring for grandchildren and their own younger children, though census transcription listed James as aged 24 and Emma as 48. Emma was a sister of Sarah Carrier, in whose home the gunfight would occur, and, forming a double bond, James was also the brother of Sarah’s absent husband Haywood Carrier. “Two sisters married two brothers,” as the family said. In the census list a son, “Carrier, Wood?” was in fact Wade Carrier, who by 1923 would be married, leaving the parental home and moving a mile up the tracks to Wylly. Meanwhile, “Carrier, Eddie” is correctly noted by the census, as is “Carrier, Willard” (a grown son with an undiagnosed mental disability sounding a bit like autism, and giving him a nickname, “Big Baby”).

Continuing with household #113, the census has “Carrier, Goldie,” nearly correct. And “Carrier, Louerie?” is son Lonnie, known as Loamy. But “Carrier, M.” apparently refers to James’ and Emma’s youngest son, little J. C., while below that, the line filled only with a forlorn question mark would seem to designate Ruben Mitchell, one of two grandchildren in the house. Next is Ruben’s sister, easier to puzzle out under the listed name “Longly, Mae Lee,” for her name in fact was Minnie Lee Mitchell. However, the “Longly” mistake, made not by an enumerator’s haste in 1923 but by transcription haste in 1993, raises deeper questions. Minnie Lee Mitchell would indeed grow up to marry and be Minnie Lee Langley. She would be the much-publicized key witness in the 1990s Rosewood claims case. But how could the 1923 census possibly have looked decades ahead to see her married name, Longly/Langley? The “University Team” transcription seems to move unconsciously like a Ouija board, superimposing the known onto the unknown.

Closely related to James and Emma Carrier's household is the next one, #114, containing the young family of James and Emma’s grown daughter Beulah, along with her husband Frank Sherman and their two small children. Beulah’s father James had built for them a small shanty next to his own larger home. The census makes an interesting stab at the name “Beulah,” calling her “Sherman, Beauty,” though this was not even her nickname. The family knew her as “Scrappy” (similarly, her tomboy niece Minnie called herself “fighty”). By whatever name, Scrappy/Beulah would be one of the heroes of the 1923 disaster.

By that point Beulah was not living next to her parents in Rosewood but was in Wylly. Her devout parents had approved of Frank Sherman as a husband for Beulah because he was a preacher, but the marriage had dissolved, as Sherman left the scene and Scrappy—living up to the resourcefulness in her nickname—took the two children to a new home in the Wylly turpentine camp. Her resourcefulness horrified her parents, however, because she became a live-in waitress in Wylly’s jook, its Saturday night dance hall. Notwithstanding this loss of respectability, it was Scrappy who would rise to the occasion on January 4, 1923, when her young relatives fled from the Thursday night shootout. Scrappy hid them in a pine grove in Wylly and kept the shivering children warm through a miserable night, when it was cold enough for frost to form. In 1994 Scrappy’s niece, by then a widow named Minnie Lee Langley, would keep a legislative hearing room spellbound with her description of that night. It was a decade earlier, in May 1983, when I originally managed to locate Minnie, for a segment on 60 Minutes. In the intensity of her story we talked past nightfall in her home, failing to notice in our concentration that no one had turned on the lights, and the reporter’s notebook I was scribbling in was almost too dark to see.

The next listed household, #115, was more than a hundred yards up the railroad from James and Emma Carrier, and on the other side of the tracks, the south side, with a small back porch facing toward Gulf Hammock. This was the modest home of James's and Emma’s oldest son, Aaron Carrier, and his wife Mahulda. The census made another brave stab at Mahulda’s name, calling her something like “Carrier, Wilermena,” but this was not even Mahulda’s nickname or middle name (which was Gussy). This household was almost the last one in Rosewood along the northeasterly course of the railroad, before open country led to the work quarters in Wylly. And this household, too, held a fateful place in the violence of 1923.

In the initial Rosewood manhunt that sparked a chain reaction on January 1, 1923, an excited tracking dog led a posse through Rosewood without turning aside, but then confidently ran up to Aaron Carrier’s front door. Yet at the same time, the posse felt confident that Aaron Carrier himself was not who the dog was following. No one was inside the house. When the door was thrown open the dog seemed to trace the movements of a fugitive who had fled through the house and out the back door—where the dog suddenly grew confused and the trail disappeared. This mystery—along with a succession of others—would set off the week of disaster.

In the census list, the next household is something of an anti-climax after the Aaron Carrier riddle, for it was not featured in the violence in any way that anyone recalled. Household #116, “Davis Hardy,” had only one occupant, hermit-like Hardee Davis, who was the brother of Rosewood’s Mary Ann Hall, the onetime young adoptee who around 1900 had gone to live in the home of then-prospering Bacchus Hall, and had begun having children there. The small cabin of Hardee Davis, like that of Laura Jones, was another place where hunters with unwanted opossums in their kill sacks might unload.

And at this household the area commonly known as Rosewood reached its eastern limit. The next home on the census list, that of Joe Robinson (#117, “Robinson, James”) was considered not to be in Rosewood but in the outskirts of Wylly. Joe Robinson was a step-brother of Sarah Carrier, at whose house the central Rosewood shootout would occur, but no rioters seeking to wreak vengeance for that shootout seem to have disturbed Robinson. Their focus, however warped, was on Rosewood.

By arriving at this figurative boundary we have accumulated a (badly mauled) numerical snapshot of Rosewood via census house numbers, starting on the Sumner side, on the west, with household #92 and ending with #116, including white households along with their more numerous African American neighbors. And as we have seen, occasionally the count is spoiled by looping out into wooded areas not generally classifiable as Rosewood, not to mention numbering errors and other puzzes.

One of the knottiest riddles is why some households were missed, not only the obscure cabin of Virginia Smith, but, especially, two homes that would be central to the coming 1923 violence. The first of these two omitted households, not mentioned here until now, was over on the western edge of Rosewood and stood near the 80-acre homestead of Cornelia Carter, who was censused as household #92. At some distance from her two-story home was a smaller house facing a road that led north, known as Highway 13.

That was the home of a notoriously irritable blacksmith and ex-convict, Cornelia’s 43-year-old son Sam Carter. The manhunt excitement of January 1 would climax, after puzzling turns, by bringing Sam Carter’s death—in what was not exactly a classic lynching but close enough: an impulse murder by a lone drunk before a provocative crowd. Sam Carter’s absence from the census seems to be explained by mobility. Once before, in 1910, a census had found him temporarily relocated from Rosewood to a town called Ellzey. For various reasosn it seems he may have gone back there around 1920, before returning to Rosewood prior to 1923. He was certainly present, and was universally viewed as living in Rosewood, on January 1, 1923.

A similar process may explain the other omitted household, the more central one, belonging to Sarah Carrier. It could be that in 1920 her home was temporarily vacant. Her husband Haywood was often away, for reasons variosly explained, and Sarah Carrier's grown son Sylvester, who in 1923 would be living with her in the house, was in 1920 living in the Sumner sawmill quarters. But Sarah's two youngest children, ages 14 and 20, were still living with her, so she could not easily have closed down the house and temporarily moved elsewhere. She had a grown daughter in Otter Creek, three of whose children were by 1923 living with Sarah. The household interactions in this case are obscure. All of this reminds that the roller coaster census tract in 1920 cannot be taken as a perfect guide. It is most useful (if combined with enough background data from informants and documents) as a general confirmation of the community’s size and condition.

The unlabeled and unexplained 72 pages of census transcription loaded into the 1993 appendix of the “university team” report to the Florida Legislature could easily give to the unwary an impression that this entire, eye-blurring mass represents Rosewood. The lack of labeling on the pages grows especially misleading as the list moves beyond Rosewood’s eastern edge and continues into Wylly. The sizeable turpentine camp there is also part of this enumeration tract, filling transcription pages 58, 60, 62, and 64 with the names of African American laborers boarding in shanties, but without relevance to the 1923 violence. In the midst of those names is household #130, “McKenzie, D. P.,” a white. Daniel P. McKenzie was owner of M & M Naval Stores, the Wylly turpentine operation. In 1923 McKenzie, along with other locally influential whites, was part of the secretive group that sought to circumvent the white rioters of early January and help Rosewood residents escape—though this was a dismal accommodation, like many of the backroom moves of Jim Crow, because it also meant leaving the physical structure of the Rosewood community, the houses and other buildings, to the mercies of the mob, writing them off as being unprotectable.

And still the census has one more surprise. Even after tirelessly knocking at each shanty door in Wylly, census taker Dorsett yet again loops back, censusing a number of white households, and a few black ones, that lie slightly to the north of both Rosewood and Wylly. And only now, as if in a crescendo, does the census finally acknowledge Rosewood’s single largest African American landholding, the 114 acres of John McCoy, making McCoy and his wife Mary Ella, the Rosewood midwife, household #144. The McCoys were linked to the 1890s arrival of the Goins turpentine business in Section 30, the light-skinned half of Rosewood, but their home may have been far enough north to cause a special census trip. Also isolated in that direction was a family of three named Borden or Barton, a mother and two girls known to other children as Cush and Punkin, about whom practically nothing has been discovered. They and a lone, cryptic householder in those woods that the census identified as “Teny, Lara” (household #145) remain to tempt further research. Unfortunately, however, the four decades that have passed since the Rosewood secret was unearthed suggest that little real research will occur, but instead alternating waves of historiographical avoidance of the Rosewood evidence and crusading publicity bursts that mostly repeat rumors and myths—as if local history in booming, fantasizing Florida is fated to remain at the level of the old Fountain of Youth, with Rosewood as one more tarted-up roadside attraction.

__________________________________

Minnie Lee Mitchell, age nine in January 1923, did not see the very beginning of the week-long Rosewood eruption, for that beginning took place three miles from where she lived in Rosewood.

But within hours the first signs of crowd excitement would come closer, moving past Minnie’s front door. Through the week as a whole, she would become the most centrally positioned eyewitness to be found and interviewed in a later era. Others in the growing informant pool had important information, and some had watched from insightful adult perspectives in 1923, but fate—on a Thursday night, January 4—would place Minnie Mitchell much more intimately and traumatically in a front row seat, confirmably in the eye of the storm.

And thus, her testimony faces legitimate questions about childhood memory. Owing to the eventual size of the Rosewood informant pool, and with help from background documents such as vital statistics records, Rosewood’s threads of suspicious-sounding testimony could be tested by what might be called consensus (i.e., if everybody else says John went to the store, but one voice boasts that he went to New York…).

Then if further probing raises larger and larger questions about an informant, emerging spots of probable misrepresentation can be set aside and not used as building blocks for an overall narrative—not finally rejected smugly, but not used for temptingly colorful details, which might have been imagined. This seldom had to be done with Minnie (Mitchell) Langley.

Among the Rosewood informants, the age of seven was repeatedly shown to be the threshold for dangers of early childhood pseudo-memories—that is, fantasies mistaken for recollections of real-life experience. White informant Elmer Johnson, remembering back to age five, was a psychological riddle in himself—reminding of the French psychologist Jean Piaget, who wrote that he grew up remembering a terrifying kidnapping attempt when he was very small, only to discover in adulthood that he was “remembering” a tale told to him in childhood by a nurse.

African American eyewitness Lee Ruth Davis, aged seven in 1923, was very cooperative as an informant, but turned out to be cooperatively promoting cherished fantasies that had replaced real memories—while she had other memories, real ones, that remained persuasively intact, and mixed unpredictably with the impostors.

However, memories from age eight and especially age nine could test out solidly, and sometimes were strikingly detailed, as if formed by an adult. In her seventies with me, and again in her eighties in the 1990s Rosewood claims case (when I also re-interviewed her), Minnie Lee Langley proved able to recall her childhood as Minnie Lee Mitchell with consistent fidelity. As with most of us, her recall was far from perfect, and occasional omissions or conflations suggested unspoken vulnerabil-ities. But her temperament itself, an offhand, communicative frankness, seemed motivated not by self-aggrandizement but simply by the satisfac-tion of doing the best she could to tell what she saw.

As contract investigator for 60 Minutes on Rosewood, I located Minnie in 1983. Before that I had been tracing Rosewood survivors for a year. When I knocked at the door in Minnie’s carport, then began explaining my unusual mission, the cautious figure in the shadows behind the screen door did not immediately unlatch it.

But then I said the word “Rosewood.” There was a sharp intake of breath from the form behind the screen, and a stifled sob. The screen door eased open. The hours-long conversation that ensued was so intense that neither of us registered the fact that night had fallen as we talked, and I was scribbling in my reporter’s notebook in the dark.

At age nine Minnie Mitchell had lived in her grandparents’ Rosewood home, facing a remote branch of the Seaboard Air Line Railway, on the northwest side (the tracks ran northeast to southwest).

Down those tracks, to the southwest of Minnie's customary vantage point and less than a quarter mile away, another home could be made out vaguely—“within hollering distance.” It was on the other side of the tracks and closer to the old center of Rosewood, in a mass of yard shrubs and subtropical cabbage palms. That was the home of Minnie’s great-aunt, Sarah Carrier, age 50. The two homes, of grandmother Emma Carrier where Minnie lived, and of Emma’s younger sister Sarah down the tracks, would form a tragic axis for the week of Rosewood violence.

The surnames require some explaining. Both sisters were married, yet they shared a last name. This was because in the nineteenth century the two sisters, Emma Lewis and Sarah Lewis, had married two brothers, James and Haywood Carrier. The two hearths were doubly bound, and in some ways were quite similar, but in others strikingly opposite.

Sarah and Haywood were extroverted performers, she a pianist in church, a loquacious organizer of community picnics; he a fiery lay preacher, in between logging work and other jobs. But Sarah’s sister Emma—Minnie’s grandmother up the tracks—was painfully shy in public, even as she was tireless and disciplined at home. Emma’s husband James, a physically imposing trapper and hunter, had language difficulties to the point of speech impediment. One of their grown children suffered what appeared to be autism, while another was known for incapacitating panic attacks. Adding to James’s difficulties recently were two cerebral strokes, just prior to 1923, degrading his speech still further. This household, loving but burdened, receded behind the talkative dominance at Aunt Sarah’s. Despite stature (Minnie’s grand-mother Emma was tall and willowy, while younger Sarah was short and stocky), and surroundings (Emma’s home was larger, more prosperous-looking and better maintained than Sarah’s), there was a balance here that would tip on a fateful week.

Both homes, beyond any possibility of exaggeration, were structures of two stories, though with a distinction. Sarah’s upper story was one large room. an attic-like sleeping area for children, without room partitions, though reached by a full-sized stairway and not just a ladder. Emma’s larger home, on the other hand, was fully built out, with partitioned upstairs bedrooms. One of the chores hated most by tomboyish little Minnie was making the upstairs beds. The sheet corners just would not seem to fall right. Minnie’s makeshift alignments sat poorly with Emma Carrier’s watchful eye.

Minnie nonetheless adored her tall, strict grandmother Emma, the only mother she had ever known. She referred routinely to Emma and husband James as Mama and Papa. Down the tracks at Aunt Sarah’s there were also grandchildren in residence, and they, too, knew their grandmother, Sarah Carrier, as “Mama.” At least a quarter of the homes that remained in shrinking Rosewood held such adoptees, as parents had died, or, in the case of some fathers, had drifted away.

Minnie’s real mother, Daisy Carrier Mitchell, failed to survive Minnie’s birth. Daisy’s husband Theron then seemed to lose interest. Minnie and her older brother Ruben became permanent wards of their grandparents.

As 1923 arrived, three of James's and Emma’s grown children had already married and moved out, but five others were still living in the household, ranging in age from near adulthood to pre-school age. With the two adopted grandchildren in the mix, it made a household of nine.

This number would have dismal significance in the wildfire rumors of a Thursday night.

____________________________________

The 1993 Report

The $50,000 academic report on Rosewood that was produced in 1993 for the Florida Legislature has in more recent years been used repeatedly by retrospectives seeking an authoritative source on the 1923 events.

But unfortunately, this has brought such distortion to the general picture of Rosewood that an issue should be confronted bluntly: The 1993 report should never be used as a source, because its solemn-sounding academic language is stretched over a daunting sea of errors. As an example of its style:

“At Lenin [probably Lucans], another hamlet located between Rosewood and Cedar Key, nine-year-old Lillie Burns and various family members watched the proceedings.” (page 52)

The brackets are in the original, neatly marking one of the report’s dismal traits, its repeatedly correcting local informants who say something accurate that does not fit the report writer’s inaccurate notions.

Here, Rosewood informant Lillie Burns Washington has accurately named the now-defunct town of Lennon, but the pompous listener, unable to find Lennon on a modern map, tut-tuts pedantically that the stupid informant must mean “Lucans”—which is the report writer’s misspelling of “Lukens,” which is in quite another part of Levy County.

All at once, the report gets a name wrong that was offered correctly, moves the location of described events far away to an irrelevant site, and then gets that site wrong, too, both in spelling and location.

The larger pattern is two-fold: 1) a remarkable ignorance of local context that is required to understand the Rosewood events, and 2) the eerie pomposity, deciding that the ignorant investigator must be right, so the facts can be squeezed to fit.

If this sounds too harsh, consider another sign of whimsy in the report, a paragraph so astonishing that it would seem surely to be a joke, if Florida had not paid $50,000 for this unintentionally humorous report:

“Elsie Collins Campbell, a white woman of Cedar Key, once lived at Rosewood, and was about three years old at the time of the disturbance. She remembered the village as one of green forests. The view is shared universally by blacks and whites when they describe the community’s dominant features.” (page 20)

Why did the report use this miraculous informant—who is able to remember back past age three, who describes a place where she didn’t even live at age three, and who is white, used for the report’s only physical description of an African American community?

The joke gets better. There is the content of this age-three memory, a memory so important that it’s being showcased. The only content of the memory is this: Rosewood had forests. And they were green. The word “green” is squeezed in to show we’re really getting detailed here. We know what color the forests were. Nor does it stop with this profundity.

The same pompous report voice expects us to sit still while it informs us, in a howling caricature of corroboration, that the greenness of the forests has been incontrovertibly proven. Why, everyone agrees that there were forests. And that they were green. “The view is shared universally by…” (Whew, we’ve now established that the forests weren’t orange or pink).

Nor is all this incidental. This report, assigned by the legislature to personnel of the Florida University System, provided the purported historical basis for the Rosewood claims case. As such, it has become the pseudo-academic 600-pound gorilla that is used credulously by journalists and other hurried scribes to depict Florida’s landmark atrocity.

Hence this one report, with its sequels, has managed the herculean task of causing Rosewood to remain in a fog of half-truths and fantasies in general public discourse.

How this happened is hinted by the Elsie Campbell paragraph’s last four words: “the community’s dominant features.”

In a rush of political urgency in the spring of 1993, the then-failing Rosewood reparations claim inspired a legislative compromise by which learned experts would supposedly establish for the legislature a profile of the contested community. The result was a chaos of academic jockeying and fumbles that might seem unbelievable if encountered in fiction. A core goal of the report, that of establishing “the community’s dominant features,” is trumpeted in the Elsie paragraph as having been successfully achieved, by means of only the green forests line, with no further description. There was no getting back that $50,000.

I know this scramble from an unusual perspective. In April 1993, as claims-case publicity was heating up, I began receiving collaboration inquiries from a Florida graduate student (letters neatly typewritten and signed, on April 5th, 9th, 30th, May 4th, and July 2nd). Far from being some kind of inside whistleblower, this remarkable correspondent, who had impressive family connections, would rise to be the moving force behind the debacle that was to be the “university team” report, finally presented on December 22.

Originally, the study had been proposed logically as a limited project to be done by the University of Florida, the closest state university to the Rosewood site, only 45 miles away. But the graduate student in question, enrolled at competing Florida State University (140 miles from the Rosewood site), proudly volunteered to me how he had used some interesting legislative rapport to get the original plan changed—so that he, a self-proclaimed Rosewood expert, could also be on the now-expanded team of five members, meaning four full professors plus his own puzzling presence.

"Our meeting with Al Lawson went much better than I had expected,” he wrote to me proudly on April 30, 1993, referring to the key legislative sponsor of the Rosewood claim. “He has agreed to change the bill and make Florida State University the primary research institution to conduct the 'official' investigation.”

This was quite a coup, apparently accomplished in part by showing around a research paper that the crusading lobbyist had written.

The paper was a wonderland of credulous fact errors, but apparently the claim’s other legislative sponsor, Rep. Miguel De Grandy, saw only its dense forest of academic-looking footnotes. De Grandy wrote to the Speaker of the Florida House:

“I am sending you an in-depth report …[whose] footnotes reference many other sources of information available throughout the state.”

Within seven months the resulting “study team” would be exclaiming in despair that there are no such written sources, but by then it was too late. The doomed study had to be slapped together for a year-end legislative deadline, using Elsie, “Lucans” and more damaging fudges.

Claims-case euphoria would then endorse the emperor’s new clothes. The original fantasy-filled paper that had impressed Representative De Grandy was never published, but seemed to hook the legislative sponsor into visions of a quick solution.

De Grandy said of the paper: “Certainly, the voluminous evidence already available will ensure that a study can be conducted in an abbreviated period of time, saving the state money…thereby providing a factual basis…”

In reality, when the “abbreviated period of time” was up, there was only chaos in the study team. Among the remarkable number of improvisations used to help willing legislative recipients agree that this was a real study, with reliable facts, a typist was paid $1,000 just to type up 72 unlabeled pages of the 1920 federal census, indecipherable by mere mortals, but journalistic enthusiasts plumping the claims case could then marvel that this monumental report was over 500 pages long. In truth there was a descriptive core of only 39 pages on Rosewood, inflated by many Elsies. The result looked just convincing enough to short-circuit any future Rosewood historiography, perhaps permanently, by throwing a spoor of false facts to be used glibly by future seekers of a quick quote.

Another jiu-jitsu move used to snatch the study plum from the University of Florida and deliver it to the activist graduate student was political correctness. The switch was covered by saying that balance also required the participation of a historically black university (Florida A & M). The resulting gerrymandered team wound up with two African American and three white members, three of the total coming from just one university, the once containing the energetic (and white) graduate student. Named as titular head was a team member both black and female, but the prominent force on the research end, who apparently did much or all of the actual writing on Rosewood, was the graduate student.

The academic seriousness of all this was suggested by that paragon’s unsolicited letter to me on April 30.

In a boast about defeating the hated University of Florida, the FSU enthusiast wrote (and did not hesitate to put it in print over his signature): “I get a great deal of satisfaction from screwing the gators.”

The resulting barrage of inaccuracies in the report, with their academic cachet, might not have been so damaging to the historical record had they been limited to epidemic misspellings and term-paper stuffings like Elsie. But unfortunately there was a central theme of fact distortion, with one aim. Team members spoke openly about their determination to help the Rosewood reparations claim succeed. Whether unconsciously or more pointedly, their picture of historical events warped in that direction.

Just after Elsie’s green forests came an extraordinary misrepresentation of Rosewood’s population figures, making the community look much larger than it actually was. This supported the “thriving” communal success fantasy that was being promoted by the claims case, and presented that fantasy as history. Hardball politics seemed to mean that just being an atrocity was not enough; the Rosewood atrocity had to be focused on a miraculous Oz. The moral nuances of all this can be debated by advocates, but it was curtains for factual history.

The team’s population misstatement was so glaring that an embarrassed team member was forced to retract it in print, though the graduate student, where it evidently originated, seemed unrepentant, later ignoring the retraction and republishing the fudge as fact in a 1996 academic paper. That paper would comprise a second wave of damage to the historical picture, also to be solemnly cited by enthusiasts in future decades. By such means a Rosewood Never-Never Land has come to float hazily but stubbornly on the Web, supplanting the real history and largely impregnable in its wreath of professorial sourcing.

The last I personally heard of the hero of this story, the graduate student, was from disbursement clerks at a state government office, where, it seems, payment of his $8,000 share in the $50,000 study had somehow been delayed. He gained notice from the clerks, they said, by shouting to the rafters about the injustice being done to him, and continued to exclaim about it outside, after walking out the door. Though this particular participant was white, the racially-focused emotions tapped by the Rosewood claims case had a way of attracting volcanic unresolved personal problems in either race.

The Rosewood claim of the 1990s had been devised in the first place, in 1991, by a tabloid TV promoter (also white) who had graduated from stealing hubcaps to armed robbery and jailhouse protests. Then the hardball politics led to that player being kicked out of his own project by another contender (this time black), who took over the publicity aspect of the case. This new aspirant also had a background as a convicted felon, for waving a gun as he threatened to kill his ex-wife’s parents. Rosewood became a kind of high-stakes, eye-gouging way for elusive dysfunction to triumphantly prove that the problem was with the system.

As to the 1993 university-team report, by October 1993 it had internally crashed and burned so thoroughly that just before deadline the team had to sheepishly call and commission me to write a $2,000 summary of what actually happened at Rosewood. So the factual issues you’re seeing here don’t come from some outland; their authority has been conceded, so to speak, by the team itself. The difference now is that the public gets to see. I don’t know whose cut my $2,000 came out of.

The many efforts in the 1993 report to fudge facts in order to pander to comforting or remunerative visions remain on the Web without a hint of the pitfalls they hide. They form a permanent temptation for anyone also wanting to fudge those facts but wishing to take shelter by attributing the fudges to a lofty voice, whether or not that qualifies as “screwing the gators.”

___________________

"They Just Want the Movie"

Prestigious Fact Errors

Other than my own early errors on Rosewood in 1982 (nearly as numerous as they were regrettable), many of today’s published Rosewood fact errors seem to track back to two main sources, both linked to the Rosewood reparation claims case in the Florida Legislature in 1991-1994.

These two sources are doubly hazardous for hurried seekers of a quick quote, whether in journalism, the blogosphere or advocacy, because superficially they look impressive to the point of stern academic monotony: dry, pedantic language, wreaths of small-print footnotes, and seeming approval by top authorities.

To the new inquirer, it might easily seem unlikely that all this could be born of a pressured moment in Florida politics, which produced an academic-looking mirage.

To get to know the two misleading sources in detail, while comparing them to the well-established but oft-buried facts they claim to represent, is to meet the rite of passage that I met in the 1990s. I learned a hard lesson then about the rigors of academic fact-checking, which I had long held in awe. It turned out that under certain circumstances there may be only a superficial appearance of such rigor, hiding a sham. In the claims case, the circumstances came when a highly emotional issue—race relations—met political excitement and pressure.

I’ll name the two sources below, then itemize a few of the main fact errors in them that tend to get repeated credulously in present-day discussions of Rosewood.

The two sources are:

1) A Documented History of the Incident Which

Occurred at Rosewood, Florida, in January 1923, submitted to the Florida Board of Regents on December 22, 1993. This is the $50,000 study that was mandated by the Florida Legislature [See the previous section for the process that produced it].

2) “Rosewood, Florida: The Destruction of an

African American Community,” The Histor-

ian, Spring 1996, Vol. 58, No, 3, pp. 605-622. This academic journal article derives from the 1993 study, using some of the same wording and information, and is authored by the same graduate student who was at the core of the 1993 study. Fact errors in this article were not only carried over from the 1993 study but in at least one case a correction that the study had tried to make to its own error was discarded by the 1996 work and the error was reinstated as fact (on Rosewood’s alleged population figures, below). This 1996 paper has often been used and cited as a source by sub-sequent discussion of Rosewood that does not consult the primary evidence.

(The errors described below, in both 1993 and 1996, have to do with inflating Rosewood’s size or the severity of its 1923 violence, promotng a picture in line with the reparation claims case. As politics, the morality of this can be debated; as purported history the result was fakery, whether unconscious or intentional.) .

Error #1: Rosewood’s African American population in 1923 was more than 300 (the real number was less than 100).

The 1993 academic study says on page 20: “Rosewood and nearby Sumner constituted a precinct of 307 people in 1910 (158 whites, 128 blacks, and 21 mulattoes). By 1920 this population had more than doubled to 638, except now blacks were a majority with 344 people, while white residents numbered 294. The Rosewood voting precinct in 1920 had 355 African Americans.” On page 22 of the study the implication is spelled out: “The village’s largest total population was seven hundred in 1915; in 1923 blacks made up the majority.” The 700 population figure is flatly false, as the study team eventually admitted, but the admission was not included in the report. As to how the figure arose, see below.

In 1996, the referenced journal article said

(pp. 609-610): “By 1915 the Rosewood vot-ing precinct had an African American population of 355 (See Table 1)…Although by 1920 the black population had declined slightly from its 1915 figure, the sense of community was still strong, and a number of small, black-owned business enterprises, including a general store and a sugar mill, continued to operate…

TABLE 1 .

State of Florida Census—Levy County

1885 Precinct—Rosewood Settlement

White: 140

Black: 84

Total: 224

1915 Precinct—Rosewood

White: 345

Black: 355

Total: 700”

[End 1996 article quotation]

First, one sees that the 1993 report chose to use the federal census, taken at the opening of every decade (1910, 1920, etc.) while the 1996 article chose to use Florida’s state census, taken at the half-decade mark (1885, 1895, 1905, 1915, etc.). That difference is not important here, for both reports used the census figures in the same misleading way. The 1996 version repeats the main illusion, leaving a good-faith reader to conclude that in 1915 Rosewood had 700 people, including 355 African Americans. It didn’t. The article doesn’t explain (as the 1993 report did try to do) that “Precinct—Rosewood” was an overall name, and that the largest settlement in that precinct was not Rosewood but the sawmill company town of Sumner, three miles from Rosewood, where the great majority of the 700 people lived.

The 1993 report was phrased in a manner that allowed the reader to assume a Rosewood population in keeping with reparation claims that Rosewood was a magically thriving entrepreneurial hub (the 1996 article’s “black-owned business enterprises” didn't exist either in 1923).

A mob attack on any community, with or without business vitality, would seem atrocity enough, but for purposes of the claims case a mirage was floated. In 1993 the obviousness of this misimpression eventually forced admission in print that Rosewood was not as large as portrayed. But in another three years the undiscouraged author of much of that original information seemed intent on it, and renewed the mirage. The 1996 phrasing, “Precinct—Rosewood,” with no qualifiers any-where, leaves the reader to naturally think that 700 people must have lived in Rosewood itself in 1915. Then allegedly this figure only “declined slightly” by 1920—which would seem to leave more than three hundred people of color still living in Rosewood in 1923, and thus inflating the picture of the 1923 destruction. In reality, survivor testimony collated with the 1920 census and other indicators leaves no doubt that the African American population of Rosewood when the 1923 violence struck was less than 100. The inflated picture does not come from survivors, but from academic enthusiasm.

In comparison to a wide world of atrocities, all of these population numbers, 700 or 100, would seem very small, almost minuscule.

And yet they take on significance by showing how emotions surrounding race relations can subvert even credentialed authorities on history.

Error #2: Inflating the Rosewood death toll.

The 1993 academic study says on page 56:

“Based on contemporary evidence there were eight deaths, six blacks and two whites.” And similarly on its page of conclusions (p. 87): “[W]e can document that eight people were killed during the racial violence—six blacks and two whites.” These numbers are accurate, but the report goes on to question its own assertion.

It follows the hedged death toll sentence with a lengthy showcasing of stories from two white informants (page 56) who I originally found in 1982, and who claimed higher numbers. Elsewhere in the pages here, both testimonies are extensively debunked, as empty boasting in one case and in the other case possible pseudo-memory after long struggles with nervous impairment. The possible causes of the confabulations matter less than the voluminous evidence against them.

As only a small sample, one of the accounts says that 27 bodies were dumped into large furrows dug by a fire plow and then immediately covered over. The other reports coming the day after that and seeing an open pit, not covered over and dug with shovels, and supposedly containing 18 bodies, not 27.

The two informants cross-refute. They can’t both be reporting facts, and indeed neither was. As said, there is much more evidence than this against the two stories.

The 1993 report’s placement of these mis-leading accounts after its weak mention of a death toll of eight ("...We can document...") has led subsequent journalists and others to intone that “eyewitnesses say” the death toll was much higher than eight.

Wikipedia goes still farther, somehow in-flating the two dissenting informants into “several,” as it declares (browsed Sept. 2022): “Several eyewitnesses claim to have seen a mass grave which was filled with the bodies of black people.” No they didn’t.

Rosewood’s complex evidence involving witness memory and racial passions has been a Waterloo for a subset of Wikipedia editors, who seem to blindly trust whatever an official-looking secondary source has written down on paper. The “several” fudge, however, was Wikipedia’s own, no source required, as it distorted even the distortions.

Meanwhile the 1996 journal article (p. 618) condenses the 1993 misdirection on the death toll into much fewer words: “The death of six blacks can be documented.”

And in this article the misdirection takes a slightly different form: “Local accounts by Cedar Key’s residents are varied, with estimates of the number of blacks killed ranging from 6 to 27.”

This at least does not endorse higher death toll estimates that are even worse illusions, but it is still false in several ways.

First, the high-end figure it gives, 27, does not come from “Cedar Key’s residents” but from a lone informant who did not live in Cedar Key, Jason McElveen (one of the two white informants discredited above).

And as far as local “estimates” were concerned, that realm of tall tales actually ranged all the way up to 150, and even 200, with no evidence or credibility; the 1996 paper makes local yarn-spinners sound more careful than they were.

As to the paper’s low-end figure, six, it is not an “estimate” like the alleged 27, but is the documented number of black dead, as the passage itself admits elsewhere.

The end result of this 1996 exercise is similar to that in the 1993 report: It has prompted subsequent journalists, Wikiped-ians, and others into claims that “eyewitnesses say” the Rosewood death toll was up to 27.

Error #3: Black-owned business enterprises.

The 1993 academic study says on page 22 that in 1923 as the mob violence arrived in Rosewood, “a number of black-owned businesses continued to operate. There was a general store operated by a white family and another by a black family. One black operated a sugar mill.”

The 1996 journal article (p. 610) slightly rearranges this phrasing, saying that in 1923 as the mob violence arrived, “a number of small, black-owned business enterprises, including a general store and a sugar mill, continued to operate.”

These references, whether to “black-owned business enterprises,” a black-owned general store, or a “sugar mill,” ostensibly showed business vitality that was destroyed in 1923.

But in reality, all are referring to the activities of the same man, Bacchus Hall, and he was dead before 1923, leaving his family in near-starvation, with no businesses at all. Hall’s only business in better times had been a store. The “sugar mill” was not a business and did not mill sugar; it was Hall’s makeshift cane press driven by a mule. His store, C. B. Hall & Co. General Merchandise, had closed down perhaps as early as 1915, due to unpaid debts incurred by neighbors, according to Hall’s son Sam, interviewed in 1983.

Farther back in time, around 1905 to 1910, there had been other black-owned businesses in Rosewood, mostly small stores but including a larger turpentine distillery. All had fallen to the shifting fortunes of a timber town, in ways that predated the 1923 violence and were unconnected to it.

The 1993 report strained so intensely to see a vision of 1923 business activity that it missed a witness statement in its own pages (p. 259, report appendix), reiterating that Bacchus Hall’s store, already closed and empty, had burned down accidentally prior to 1923.

Error #4: Did Rosewood’s history as a mostly African American community date back to 1900?

The 1993 academic study says, on pages 21-22: “By 1890 the red cedar had been cut out, forcing the closing of the pencil mills at Cedar Key. The community had a black majority by 1900, as white families moved out, leasing or selling their land to blacks. The post office and the school closed, relocating to the site of a new cypress mill that opened in Sumner.

The 1996 journal article changes the wording slightly (p. 609):

“By 1890 all the red cedar in the

surrounding areas had been harvested and the Cedar Key pencil mills closed, persuading most white families to move out and to sell or lease their land to blacks. The post office and school closed, relocating to the site of a new cypress mill which opened in Sumner…by 1900 the black community had become the majority.”

Both in 1993 and 1996, this picture of Rosewood’s history was badly flawed. The shift to a black community is being moved back by a decade or more to give an appearance of longer-term stability. The Cedar Key pencil mills did not close until 1898, when they were destroyed by a hurricane, and Rosewood’s cedar sawmill continued operating until 1911, when at last clear-cutting had exterminated the swamp cedar. It was at that time that many whites moved out. The new population magnet was indeed a cypress sawmill in Sumner, built in 1911. Rosewood’s early history as given in both reports, 1993 and 1996, is dotted with fantasies and false assumptions. This early history is perhaps incidental to the later violence of 1923, but its imagined picture puts discussion on shifting sands. Here, too, the imaginings have been embraced and used as fact by subsequent journalism, turning history into a dreamscape.

Just in the short passage given above, the errors or misdirections are various. The Rosewood post office did not close until 1914, and “the school” apparently means only a white school, while a one-room, county-financed school for African Americans, of eight grades, continued in Rosewood despite population changes.

One thing that both of the 1990s reports were hiding was that the author (the same author in both cases) did not know which households in the census were in the area considered to be Rosewood, and so could not really determine when the shrinking remnant community came to have a black majority. This ignorance was also confessed by the 1993 team.

The above samples present only a fraction of the reasons why the two prestigious-looking reports, in 1993 and 1996, should not be used as source material on the Rosewood incident. The smaller errors in those works are too numerous to catalog. A casual inquirer, who does not know the primary evidence that the reports are distorting, has no way to know which assertions are or are not factual.

And this presents a deeper dilemma: If there has been so much fantasizing about Rosewood, including in the two academic efforts, then what authority should be consulted when people want to know about the 1923 atrocity?

Of course, they could look here, at the pages you’re reading now. But why should they? What kind of institutional imprimatur do I have? What body of authoritative opinion is there to certify that any work on Rosewood is trustworthy?

The answer may be that Rosewood, as history, is doomed permanently to ambiguity—doomed by the emotions in racial controversy and perhaps by the uniquely fantasy-prone culture of Florida.

_________________________

Bryce (and the "brothers" myth)

As Rosewood residents fled their homes on Thursday night, January 4, 1923, and the women and children sought refuge on protected ground--that is, on the property of influential local whites--the improvised sanctuaries were spread over an area ten miles long. maintained by four men: 1) storekeeper John Wright (inside Rosewood); 2) sawmill superintendent Walter Pillsbury (Sumner); 3) turpentine camp owner D. P. McKenzie (Wylly), and 4) turpentine woods rider Jack Cason (near Otter Creek).

But soon there would be a fifth man, more mobile than the rest, and forming the crucial link in a plan to remove Rosewood refugees farther from danger.

“Tickets, please. Tickets, tickets”—this cliché picture of the conductor on a passenger train was only the publicly visible part. As any railroader knows, a train conductor is captain of the ship, while the engineer is his pilot. As the engineer marks time at the throttle, the conductor is back at his papers in the caboose, toting up manifests, sifting through directives. And William Crighton “Kay” Bryce, conductor of the Seaboard passenger train out to Cedar Key, was the beaming, lordly epitome of this role. Moreover, in the adventurers’ melting pot of frontier Florida, he was perhaps the world’s most unlikely train conductor.

Coming from old family wealth, with a plantation the size of a small community at Bryceville and a turpentining operation down the line at Kaytown, he wasn’t into train conducting for the money. He had loved trains since the days when he was punished as a child for running off and hopping onto them, said his daughter Leona.

In the grim plotting of January 5-6, Kay Bryce did not operate a stationary refuge, but ran a larger mobile one that would substitute for the others. The idea was that after the passenger train left Cedar Key as usual at 5:30 a.m. on Saturday, it would make a series of secretive stops, at each of the sanctuary nodes, taking on women and children, then taking them all to safety in Gainesville. The fear surrounding this plan was apparent in its prohibitions; as in the original waystations, none of Rosewood’s fleeing men or older boys were allowed on board. The Lexie Gordon and Mingo Williams killings had demonstrated that there were men in the crowds now roaming through Rosewood who might seize any excuse to twitch the trigger.

The rescue plan, like many actions by outnumbered authority in a riot, was not a triumph but a dispiriting limitation of defeat. Families were atomized by the gender rule, and the worst defeat did not need to be spelled out. To appease the mob, giving them a token prize to crow about so that hopefully they would go home, the remaining buildings of the community were being surrendered to them. The deserted homes, barns, and other structures would stand in wan expectation through Saturday, like a condemned man at the noose, waiting for the flames that no one would stop on Sunday.

Little Minnie Mitchell, half frozen in the pine grove before dawn on Saturday, was elated as she found herself being led from hellish outdoor concealment into the smiling care of Bryce the train conductor, as he gave her a box lunch. She wondered why some of the older women coming with her didn’t seem to feel her joy and relief. Didn’t they know they were being saved? Why were they crying?

And one last thought on Kay Bryce, the improbably rich train conductor with the heart of gold. It’s unsettling to me now to listen to my informant tapes from what sounds like the Stone Age of pre-digital information, back in 1982. It’s unsettling because it reminds how lost I was in the flood of legends and fantasized myths that came swarming out when I unlocked the Rosewood secret. Writing the opening article was like batting at moths under a streetlight. I had to choose, on far too little corroborating evidence back then, which of the fervent assertions to accept as telling the real story of Rosewood.

And in that confusion, somebody, somewhere along the way, told me vaguely about two fascinating train conductors. They were brothers, the story said, and had orchestrated the dramatic rescue by train. I was told that they were John and Kay Bryce—and I verified that they existed, sure enough.

The problem was invisible to me at the time: I had it on far too little authority that both brothers had participated, together, in the Rosewood train rescue. For they didn’t. Only Kay Bryce was on the rescue train. His brother John was apparently on another run. Lamentably, in my investigative article of July 25, 1982, I attributed the train rescue to them both, as an endearing brother act.

However, there was a silver lining.

What I further failed to guess was that in later decades my fumbling error would inadvertently provide a remarkably effective litmus for testing which emergent journalists, authors, advocates, bloggers and other commentators on Rosewood were heedlessly lifting from my old work, without bothering to check. It seemed to make little difference to the lifters that I had long since corrected the original Bryce brothers error in print. They scarcely stopped for such inconveniences. After a while they didn’t even need to lift the error from my own gaffe in 1982, but could lift it from other lifters, whole generations of them, who, with similar heedlessness, seemed to like the irresistible picture of the brother act, without much caring whether it was true.

“The ordeal ended due to the efforts of two white brothers, William and John Bryce,” declared the $50,000 report prepared for the Florida Legislature in the Rosewood claims case (p. 61), amid a sea of other errors in that report.

The myth that might be called The Flying Bryce Brothers was like a circus sensation. In the claims case of the 1990s it grew so potent that still further embellishments arose. Suddenly the Bryce Brothers, who supposedly saved Rosewood together in a fantasy duet, weren’t just wealthy hobbyists. A 1992 fantasy said they owned the whole Seaboard railroad, and that these two miraculous brothers had adventurously moved to Florida from up north (the real Bryces had old southern roots going back for generations).

The Washington Post then thrilled to the discovery that the bold Bryces came from up in Post country: “Other heroes may be John and William Bryce,” announced the Post on May 39, 1993, “two wealthy white Northerners who sent a slow-moving train through Rosewood…” The Post's version of the rescue train did have a certain northern brusqueness, not deigning to make a full stop for the hustled refugees and only slowing down (unlike the real train), but the Post imprimatur called to generations of future imitators.

Still further, wide-eyed discoverers began saying, for reasons impossible to guess, that the two heroic brothers were not just northern transplants, but were Jewish (though for generations the real Bryce family had been Methodist).

On September 23, 1993, the Rosewood claims case “academic team” weighed in on The Bryce fantasy by interviewing claims case publicity liaison Arnett Doctor, he of the mansions with manicured lawns, the 30 women and children buried alive, and generous misattributions to me.

“The Bryce brothers,” Doctor revealed confidently, “were a couple of brothers out of New York that came down to Florida to make their living. They were Jewish people and they had some sort of enterprise in the Levy area. They were entrepreneurs in the Levy area.” (Now they become not train conductors but some kind of clever Semitic fixers). The description was not finished: “And it was my understanding that they had a very close relationship with the people who owned the Seaboard Railway. Supposedly there was a joint effort by Mr. Wright and the Bryce family to get the train to come in….By that time they were aware of the fact that the soldiers from Camp Blanding, Florida…came in and stood guard.” (This, on page 157 of the 1993 Rosewood academic report appendix, lumps the brother act in with some non-existent soldiers).

So how could the story of a real atrocity start to look like a slapstick farce, with mythical participants popping in and out of weighty tomes and media masterpieces? Could it happen only in ditzy Florida? Or was it the whole nation’s phenomenally fantasy-prone racial complex that was at fault?

The siren call of wonderstruck storytelling seemed to have a special hold on Rosewood—whose real facts were already so like a demonic fairy tale that artful embellishers could scarcely resist taking it the rest of the way—to create the perfect bygone paradise destroyed by the perfectly gigantic massacre—amid the heart-warming tap dance of the two northern Jewish brothers.

At any rate the point here is that if you, dear reader, should happen to come across any commentary on Rosewood that exalts the two brave Bryce brothers (rather than the one real conductor), you’ll know that you’ve been privileged to meet a gem-like indicator that you’ve stumbled onto a source not very given to real research.

_____________________________________________

The Lost Paradise Myth & Evidence Index

“Black folks own all the land 'round here.

All the businesses, too.”

—actress Esther Rolle, as Sarah Carrier

in the 1997 motion picture Rosewood

(scene chosen for especially heavy trailer promotion)

The Rosewood that was pictured in the 1997 movie is not just a myth, but is a remarkable hoax, in various ways pretending documentary authenticity.

There were no businesses at all in the Rosewood that met 1923—except for one store, whose owner was white. The majority of the land was also owned by whites.

In the scattered two-mile span (across Sections 29 and 30) that contained the hamlet of Rosewood, whites owned more than half the land area. In the part of Section 29 forming the old center of Rosewood and the focus of the 1923 attack, African American land ownership was in small parcels, sometimes an acre or two, in one case mounting to 40 acres, in another 27. African American ownership represented less than one-fifth of that section’s total 640 acres.

These real black land holdings were noteworthy tributes to hard-working Rosewood residents, but they are trivialized by the exaggerations. The recurrent claim that the destroyed community was a powerhouse of black-owned businesses is a central illusion in what might be called the Rosewood Myth. It is alleged—always vaguely, without traceable specifics—that if not for the 1923 destruction, the promising future of this imagined town could have risen in future decades to a status of urban impressiveness, leveling the racial playing field. This is the generational wealth argument, saying that Rosewood’s tragedy demonstrates, not just mob cruelty, but unfair competition by whites, on a scale that could contribute to “the historic inability of African-Americans in this country to acquire capital in one generation, and pass that capital along intact to the next generation."

The truth was that there was little capital to pass on, and in various cases what capital there was did get passed on, despite the 1923 violence. The largest Rosewood land parcels owned by African Americans (up to 80 acres and 114 acres in Section 30) were retained by those owners, not confiscated in 1923. Then by 1925 most were sold by organized effort among the Rosewood owners and heirs as prices rose in the Florida Boom (just before the boom’s 1926 collapse left the site of Rosewood in even greater isolation than before).

The quote above is from attorney Steven Hanlon, who managed the successful Rosewood reparations claim in 1992-1995. In the fall of 1993 the claims case promoted ideas that Rosewood residents were “the second largest land owners in Levy County,” supposedly owning “several thousand acres”—an incredible fantasy, which never had to face challenge in the waves of enthusiasm that accompanied the claim.

On a national scale, for the generational wealth argument to be extrapolated from Rosewood, the nation would have to be hiding thousands, or tens of thousands, of mythical Rosewoods filled with mythical businesses that were mythically destroyed. Such an adjustment would make early twentieth-century America not just racially flawed but more like the killing fields of Cambodia.

For individual Rosewood residents like Minnie Mitchell, the 1923 pogrom was a crushing disaster as it took homes, possessions and loved ones, but it was not a political melodrama about alleged black dynamism being snatched away from a promising destiny.

The real Rosewood that woke to the morning of January 1, 1923, was already at a dead end. As 1923 arrived, faint traces in spreading weeds marked the sites of former black-owned businesses that had genuinely existed in Rosewood’s small boom period of earlier years, as manic spikes in the timber industry left towns and camps winking in and out of existence.

There had been Ransom Edwards’s store, Bacchus Hall’s larger store with a soda fountain downstairs and caskets sold above, John Coleman’s store, and the two smaller “shops” of Charlie Williams and Ernest Blocker (the innocent word “shop,” in this region, was like still more innocent-sounding “grocery store”—often meaning moonshine and a back room for cards—as it apparently meant in these cases).

Ironically, hard times had cleaned Rosewood up, leaving no “jook” as in the timber towns that still bustled on either side: Sumner and Wylly. Wet or dry, survivors agreed that all of Rosewood’s black-owned stores had failed as hard-pressed residents in a dying local economy neglected their credit tabs. The vacant plank structures were left near a steam-powered railroad with its flying cinders, and did not last long against accidental fires. Hence the faint rectangles in the weeds.

But the central mirage in myths about Rosewood The Destroyed Entrepreneurial Giant was another kind of landmark, whose very nature tends to stump modern inquirers, creating puzzled guesswork as if to say, “Well, gosh, anything must have been possible back in those strange old times.”

This landmark was Rosewood's turpentine distilling company. M. Goins & Bros. Naval Stores (even its name tends to be a stumper for post-millennials) had genuinely been a thriving fixture on the western edge of Rosewood, and was genuinely a rare kind of success story: an African American-owned industrial firm that had pushed into a white-controlled area of the Jim Crow South—and flourished there.

For about a decade. The boom days for the Goins turpentine distillery were around 1895 to 1906. Turpentining—the boiling down of sweet-scented pine resin to make pitch, tar (“naval stores”), rosin and multi-use spirits of turpentine, had since colonial times been a mainstay of the coastal South. By the 1890s an aging treasure trove of pine trees in North Carolina had been largely tapped out by destructive resin cuts, which could ruin a tree as lumber.

A mass migration of tupentiners began moving south, into virgin Florida, including a family that owned 4,447 acres in the North Carolina sand hills near Fayette—land whose remainder would be sold in 1913 to become part of Ft. Bragg. Family leader Martin Goins (the “M” in M. Goins & Bros.) shepherded a transfer by rail of distillery equipment, family, skilled employees, and furnishings, including at least one piano. The new arrivals planted in forbidding Florida wilderness some well-appointed homes, a workers’ quarters, a pay office with bars on the window, and two grown daughters polished by St. Augustine College in Raleigh.

What puzzled old neighbors of the new distillery, however, was how to classify the boss. To shrugging white woodsmen he looked white, but Martin Goins reportedly insisted that he be classified as a Negro. The census defined him as “Mu”—mulatto. Two of his pale-complexioned nephews would marry into an older Rosewood family, the Carriers, whose complexions were quite dark.

The turpentiners who came with the Goins brothers were much the same, bearers of a tri-ethnic legacy in the Carolinas: part Native American, part European American, and part African American. The Goins wealth led back to a colonial frontier land baron, Duncan Murchison, who had at least two common-law wives, known colloquially as “Indian wives.” The details went unrecorded, according to assiduous family genealogist Helena Hendricks-Frye.

Generally, the niche for tri-ethnic mystery heritage in the Carolinas goes by cryptic names: Melungeon, Portagee, Black French, Black German, Brass Ankles, “Turks,” “Cubans.” Around Rosewood, some older settlers called the new arrivals “Croatans,” conjuring tales about the lost colony of Sir Walter Raleigh.

An invisible mile-long boundary, the north-south section line between Sections 29 and 30, became a novel and usually amicable color line, separating the mostly dark old settlers of Rosewood, in Section 29, from the pale, “Indian-looking” late-comers, who prior to 1900 acquired more than half of Section 30, forming parcels of 65, 80, and 114 acres..

The problem with the Rosewood Myth—which asserts a 1923 attack on a full-sized town of “businesses”—is time. M. Goins & Bros. Naval Stores was definitely an African-American-owned business, and it definitely stood within the two square miles roughly thought of as Rosewood. But turpentining was a transient operation, moving as the trees gave out. At the beginning of the Goins presence in Rosewood, an impressive seven square miles was leased (not bought) by M. Goins & Bros. between Rosewood and the Gulf, for pine resin “dipping.” But that lease expired in 1906, and only 65 acres near the section line was owned outright by a lone surviving Goins brother, Edmond. In 1900, swamp fevers had killed the firm’s main technician, brother-in-law Eli Walden. Apparent fever took Martin Goins himself in 1905.

Brother Ed Goins continued on a series of small or disputed leases in deepening debt. The 1910 census found Martin’s widow Lydia working as a “laundress.” By 1916 the firm was gone, auctioned off to creditors, though a confusing title swap enabled remaining family members to keep the 65 acres. By 1920, all the heirs but one had moved out of the realm of swamp fevers. A lone Goins household, that of Ed’s son Perry, continued living in the old home place, which faced across the Seaboard tracks toward the defunct Goins pay office, now decaying and laced with vines. Ed Goins then died in 1920. As 1923 arrived, the last Goins householder in Rosewood, Perry, was supporting his family on manual labor, stacking lumber in nearby Sumner, to which he walked each weekday morning for two miles along the railroad.

In 1982, I too fell for the illusion that a black-owned turpentine business was still operating in Rosewood when it was attacked. In that beginning period of investigation I was limited mostly to rumors on who owned what. The old records stuffed in a courthouse basement would take years more to unravel. The stubborn rural mysteries of Rosewood would come to seem specifically unsuited to the deadline pressures of journalism. Instead, they call for the patience and focus of a historian, though even that hope badly fell to pieces in the Rosewood claims case of the 1990s. Journalism up through recent years has continued faciley citing the turpentine reference from my 1982 initiation, without any attempt to update or to contact me for details—which, of course, would only make more work for the contacter, whether or not on deadline.

Florida fantasy may continue to be cooperatively supplied with story after story applauding a non-existent 1923 black-owned store, and a non-existent 1923 turpentine distillery, along with unnamed "businesses," all supposedly falling from the American skyline in a titanic attack on generational wealth.

_________________

The Fake Soldier

“NINETEEN SLAIN IN FLORIDA RACE WAR,” cried the headline on January 13, 1923, anchoring the front page of the Chicago Defender, the largest and most widely circulated newspaper for African Americans in the 1920s.

This Rosewood coverage was a bit belated, in part because the Defender, like most alternative-niche black papers of the day, was a weekly. However, this weekly circulated not just in Chicago but across the nation, especially in the Deep South, where readers could find an alternative to mainstream white-run dailies that teemed with "black brute" language, and the word "citizens" was code meaning only whites.

Carried into the far corners of Dixie through arrangements with railway porters, the Chicago Defender was passed hand to hand in barbershops, beauty parlors, stores and cafes. In Chicago it was sold by hundreds of newsboys. The paper's founder and publisher, Robert Abbott, was sometimes called the nation's first black millionaire.

A controversial figure, Abbott had insisted that two successive wives call him only “Mr. Abbott” (according to the divorce papers). Nor is his personality a side issue here, for he was known as a hands-on editor, perhaps having personal input into the Rosewood story of January 13. By whatever hands, however, the “NINETEEN SLAIN” announcement was just a warm-up.

Readers learned that the distant events in a place called Rosewood had revolved around a startling hero named Ted Cole, who, the Defender exulted, was a formidable black ex-soldier, and one day, just by coincidence, this hero was passing through the remote swamps of Florida (from his home in Chicago, where, as it happened, he was also in the glamorous movie business). Thus ex-Sergeant Cole had landed in the middle of the Rosewood violence, and he handily took charge, teaching Chicago can-do spirit to cowering rural bumpkins, and showing them how to fight back. In epic battle they then mowed down row after row of marauding whites (though only fourteen of the "NINETEEN SLAIN" were white dead, the story modestly qualified—still a tidy score for the home team).

Or in other words, the way the Defender had miraculously learned a Rosewood death toll of nineteen was that they made it up, like all the rest of this incredible, free-floating fantasy that presented itself as real news (aided by a few details lifted from the Associated Press wire for effect).

The real people who had lived in Rosewood, the survivors who were found in the 1980s, never heard of any ex-soldier named Ted Cole, or anyone like him, while the real combatant in the real Rosewood battle, the solitary figure of Sylvester Carrier, was not in the army in World War I but was in prison.

The sheer size of the mendacity in the Defender story cuts many ways. It reminds that while the white press of 1923 looks abysmal in its creation of a fantasized black uprising and other flourishes, the cloistered nature of the black press in the same era could spawn fantasies that were still worse—or at least worse in the sense of being more audaciously fictional and more cynically deceptive in masquer-ading as real news.

The story exalting “Sgt. Ted Cole” is such a pastiche that one can read down through its entire first section, line after line, and still be left wondering where the titan Cole comes in, or how the figure nineteen was calculated onto the dead (the text beneath the headline says only “14 whites were killed”):

EX-SOLDIER

IS HERO IN

BLOODY RIOT

Makes Farm Hands Fight Back

Had Just Come to Town

Called Chicagoan

By EUGENE BROWN

Tallahassee, Fla., Jan. 12—Fought to

a standstill, repulsed, defeated in his

own bailiwick, the Southern lyncher

has fled from an open attack against

Race men and women at Rosewood,

where 14 whites were killed, and re-

sorted to guerrilla warfare, sneaking

out in parties of death at night to

waylay and slay defenseless farmers

and villagers. In this manner the toll

of death has risen above last Thurs-

day night and Friday, when a band

of determined Race men and women

held at bay and then threw back

1,000 bloodthirsty Southerners bent

on teaching them a lesson…,

Only on line 34 does one meet “Ted Cole, former sergeant in a pioneer infantry regiment which saw service in France” and who “led and inspired his brothers in blood against the assaults of the murderous mob. For the first time he was able to bring a group of lowly workers, men taught to be humble, together and make them fight.”

Notable is not just the deception but the con-descension toward the real people of Rosewood, who have to have a dauntless Chicagoan come and prod them. The story dreams up imagined conversations, a fake interview, and fantasized atrocities, along with a dateline (“Tallahassee”) whose position over 100 miles from Rosewood suggests that even the intrepid reporter’s name, Eugene Brown, may be a fake.

But rising above the rest is the ex-soldier:

“He pleaded with the men to protect their women. He pleaded with the women to stand by their husbands. He made little children cry that their daddies protect them. And the men did.”

At this point in the fantasizing, more than 100 lines into the story, there is still no explanation of the nineteen dead in the headline, though the implication would be that if fourteen whites were killed, as claimed in line 6, then the other five must be black. The oratory doesn’t pause for such details.

However, the massive headline by itself, so large that it would easily be legible from across a room—“NINETEEN SLAIN IN FLORIDA RACE WAR”—may have been sufficient for distant effects. In 1982, when Rosewood survivor Philomena Doctor insisted to me that in 1923 at age 11 she had stepped over nineteen dead white men on her grandmother Sarah’s front porch, I tried fumblingly to give her the benefit of the doubt, by asking if she meant “about 19.” No, she replied angrily, there were exactly 19 corpses—which made it sound as if she somehow counted them meticulously even while fleeing in terror for a moment across a porch. Except for her next statement. There were exactly 19, she retorted, because “it came out in the paper.”

The headline in the Chicago Defender was apparently referring to a total of both black and white dead, but it was all fantasy anyway. It was not completely impossible for a well-worn issue of the Defender to have existed in Florida, perhaps having made its way to accustomed outlets by rail, then to be seen at a glance by a traumatized 11-year-old still trying to make sense of what she had endured.

But the story of the miraculous Sgt. Ted Cole was far from being finished, as Rosewood continued its swelling informational journey.

Seven decades later, busy Hollywood researchers apparently found the old Defender fantasy in its archival grave. Though the ex-soldier story might be viewed as the single biggest lie told about Rosewood in the press of 1923, it was precisely this story that was selected by the creators of the 1997 film Rosewood to use as their fictional centerpiece, in a movie disguised and promoted to look like “history” (the word used glowingly about the movie by Oprah Winfrey). The film got large boosts from idealizing reviewers, though its grating non sequiturs crashed at the box office.

The research department that tried to pull off the historical effect was undeterred. Reportedly, $1 million of the film’s production cost was spent just on its realistic-looking period sets, which presented Rosewood as a full-sized town, filled with the “businesses” imaginatively spied by professors writing the 1993 Florida Legislature’s claims case report. As the movie reached its apocalyptic climax, a documentary-style screen note drove home the impression that this relentlessly fantasized collage was supposed to be taken as being only slightly fictionalized around the edges.

Meanwhile, the fictional warrior from the Chicago Defender, Sgt. Ted Cole, was resurrected as the starring character in the film. On-screen, this elaborately copied ex-soldier character was not named “Ted Cole,” however. The creators delighted to switch his name, too, giving him a monosyllabic surname handle, “Mann.” He was called by that throughout the show, as he blazed away with six-guns from horseback, as if the ex-soldier had parachuted into a western horse opera, and then he tenderly courted a dainty, articulate schoolmarm, who was somehow named Scrappy (without explaining that the real Scrappy in the Rosewood events, while heroic in the final analysis, was a hard-luck jook waitress living in a cubicle under the dance floor).

There was an exquisiteness in the film’s desire to play God and switch around its characters and personalities, while using real people's names.

In real life, the nickname “Man” (with one “n”) really was the nickname of Sylvester Carrier, the hearty backwoods boaster and lumberjack. In the film, the unfathomably god-like rearrangers took that name, “Sylvester Carrier,” and pinned it onto a sensitive “music teacher.”

The real triumph of the unseen cinema researchers, however, was in the movie’s inexplicably vicious slander of the white storekeeper John Wright. Though all survivors agreed that Wright was a hero in the grim Rosewood events, and no one said a word against him, the film could not seem to resist making him a philanderer, a drunk, a cheat, even a child beater—and pinning his real name onto these imagined traits.

This was where the researchers were so helpful. Dimly suggested by the outcome is a studio legal department effort to find a character with no living descendants who might sue for slander over a little imaginative rearrangement. And voila! The great pain of John Wright’s life, which neighbors said he never got over, was the loss of both his small sons around 1900, when one was an infant, the other a toddler. Ever since, he was said to have behaved with exaggerated tenderness toward children, including African American children. The film, however, seemed to find a delicious opportunity to twist the knife. It resurrected the two sons and age-progressed them to 1923, where it showed the drunken, miserly, lying, cheating John Wright abusing them cruelly, even slapping one of them.

It’s not just anyone who can hate so mysteriously that they will take the most precious thing in a stranger’s life and pervert it to smear him relentlessly. Indeed, it would seem to have taken more than one mind, first in the research cubicle, then in the scripting process, then in the editing suite, to craft such a tour de force of giggling secret sadism. And no one need ever know exactly who they were.

________________________________________