edit in Google

Lone Rock Stockade and Digitizing Prison Labor Records

This dataset is records from the convict-leasing program run by the State of Tennessee at Lone Rock Stockade in Tracy City, where over 10,000 prisoners were lent to the Tennessee Iron and Coal Company between the years of 1870 and 1896.[1]  The dataset’s structure and scope is directly indicative of how prisoners were treated, managed, and tracked: emphasis was placed on minimal identification rather than any other nuance. Similarly, the details of how the digitization project began show how understudied and forgotten prison labor during the Reconstruction of the South after the Civil war is.

The archives, under the header of “Convict Records”, have 13 columns across two pages. The qualitative fields are name, color, county, crime, eye color, hair color, complexion, and marks, while the quantitative fields are the date of sentence commencement, term length, age, height, and weight. It is written in English, in cursive, and on a set template of paper with printed column headings and with spacing that is clearly specialized for the purpose of fitting each field appropriately. Every entry was written by the same hand, as evinced clearly by the perfect uniformity of the description of physical marks: every cell began with the word “scar”, making it abundantly clear. Curiously, the name of one of the entries is deliberately crossed out using a different writing utensil (much bolder and differently sized) with no indication as to why the record is being marked. There are no other notes or extra add-ons in the margins.

The records are all from roughly around the same time frame but were not organized in proper temporal order, as they are from a 4-month time frame in the summer of 1890. The dataset spans from 1870 to its end in 1896, when the practice of incarcerated labor was ended due to free workers complaining about convicts stealing their jobs and burning down the stockade.[2]  Many of the convicts who were being leased are very young, going as low as a 16-year-old being convicted on a larceny charge for 2 years. Both white and black prisoners (marked C for “colored” in the dataset) were put to work.

It is important to note that this data was not generated by the company they were being leased to, Tennessee Iron and Coal, but by the state upon their entry into the legal system. The inmates would then be passed to the company, who took on the responsibility of feeding and housing them in exchange for free, unregulated labor.[3] These records are the bare minimum. This dataset’s very limited details are a direct reflection of its purpose: to keep track of basic identifying details. All that is conveyed is the rough likeness, name, and county of provenance for hundreds of men, and then they disappear into the black hole of prison labor. One learns very little about the prisoner from these records outside of identifying characteristics. This is, of course, very telling considering that these men were convicts being leased to industrial facilities, essentially working as slaves. Interestingly, the records do not have any gender data while there is evidence that women were held in the stockade as prisoners to be leased out to members of the community to do housework.[4]Also, as time went on, the prison demographics shifted towards being predominantly black, a direct reflection of the Reconstruction’s use of discriminatory sentencing practices when it came to petty crime enforcement and prosecution.

I would be very interested to see the contracts which the state would have been signing with Tennessee Iron and Coal for each prisoner and if there were data-based stipulations on price and expectations. The most closely related data from the course is American Slavery as It Is , because if one was to write out the data in the Tracy City archives into magazine advertisement form, it would look quite similar but be even more vague.[5] Slaveowners, because they privately purchased slaves as property and had long-term incentive to maintain their health, clearly put more emphasis on identification, account, and data-gathering. The state, on the other hand, had a legal mandate to hold their prisoners and sell them off as needed without principal payment. The Tracy City prison labor records contrast to the Ellis Island Passenger Arrival Lists. The New York immigration records have far more detailed questions about provenance, financial state, abilities, and destination, while having almost no questions on visual qualities for identification. [6]Tracking immigrants and their abilities at one singular point of intake would be fantastic for a bureaucracy’s ability to predict and calculate, while visual identification would be futile in such large numbers; conversely, until the prisoners died or were released the state knew where to find the prisoners-- focusing more on identification in case of escape would utilize their limited resources more effectively.

This data set’s relatively late discovery also shows how understudied the phenomenon is: the visiting faculty member at Sewanee who launched the digitization project only heard about it through an “offhanded comment at the Grundy County Historical Society.”[7]  Considering how many people died, how much money the stockade man for the state, and the sheer scale of the prison labor, it is surprising that no other records related to the projects have surfaced for digitalization and that it was not explored prior to this curious professor: only in 2019 did excavations of the Lone Rock  Another very interesting facet of this digitization project was the discovery that it was intimately linked to the academic institution undertaking the research. The Tennessee Coal and Iron Company “donated 5,000 acres to the Episcopal Church to help start the University of the South”, and while the donation occurred before the use of prisoner labor was widespread, the link between the two had been obscured.[8]

The Tracy City prison dataset has been partially digitized but is still rather useless. It is riddled with obvious spelling mistakes, missing entries, and plainly incorrect transcriptions. This seems to be a clear downside of crowdsourcing, as people can simply do the wrong thing. The first step towards making the data usable would be to correct these errors. Some interesting applications of the records would be to properly track demographic changes over time. While the Sewanee researchers did quantify the number of black inmates compared to white over time, other factors such as age and term length could be useful. This also calls into question whether there is a digitization effort behind the scenes that might not be publicly available yet. The way in which I encountered the project and am basing my assumptions on was through From the Page, which shows the project as active and having been edited quite recently. Analysis on the prisoner’s county of origin and crime could also make for interesting studies in tandem with policy analysis on Reconstruction policing and the Black Codes in Tennessee.

To digitize this dataset, I would take the existing work done on From the Page and employ students or research team members to finish it—around half of the archives have been transcribed, but with the safe presumption that many entries might need to be double-checked, there would still be a lot of work left to do. A visualization map of where prisoners hailed from, how long they served in the mines for, and interactive charts on different charges that led to their incarcerations would be very interesting. While not an immediate part of the dataset, Sewanee’s article about the project mentions the obscene rates at which the state leased out it’s convicts: at their peak, healthy inmates were worth the equivalent of $2.9 million dollars today, per year: TCI could feed and house them for minimal overhead, provided harsh quotas under threat of beatings or whippings, and had no accountability to the state should they die or be harmed.[9] This is obviously simlair to slavery and bringing in the necessary data points to illustrate that would be invaluable. Clearly the researchers have the data, so an important next step to properly digitizing the entire project would be to bring them together into a clearer interactive picture.


Parrish, Kate. “Ghosts of Lone Rock.” The University of the South. Sewanee. Accessed November 16, 2021. https://new.sewanee.edu/features/ghosts-of-lone-rock/.

Weld, Theodore Dwight. “American Slavery As It Is: Testimony of a Thousand Witnesses.” Theodore Dwight Weld, 1803-1895. American slavery as it is: Testimony of a thousand witnesses. University of North Carolina at Chapel Hill, 2000. https://docsouth.unc.edu/neh/weld/weld.html.

[1] Parrish, Kate. “Ghosts of Lone Rock.” The University of the South. Sewanee. Accessed November 14, 2021. https://new.sewanee.edu/features/ghosts-of-lone-rock/.

[2] Ibid.

[3] Ibid.

[4] Ibid.

[5] Theodore Dwight Weld, “American Slavery As It Is: Testimony of a Thousand Witnesses,” Theodore Dwight Weld, 1803-1895. American slavery as it is: Testimony of a thousand witnesses. (University of North Carolina at Chapel Hill, 2000), https://docsouth.unc.edu/neh/weld/weld.html.

[6] “New York Passenger Arrival Lists (Ellis Island), 1892-1924.” Accessed 5 Oct. 2021. FamilySearch, 23 Feb. 2018, familysearch.org/ark:/61903/3:1:3QS7-89ND-LLHR?cc=1368704&wc=4FMB-7KL%3A1600302329.

[7] Parrish, Kate. “Ghosts of Lone Rock.” The University of the South. Sewanee. Accessed November 14, 2021. https://new.sewanee.edu/features/ghosts-of-lone-rock/.

[8] Ibid.

[9] Ibid.