From The Connecticut River Valley in Southern Vermont and New Hampshire: Historical Sketches by Lyman S. Hayes, Tuttle Co., Marble City Press, Rutland, VT., 1929, page 276:
A CURIOUS CENSUS ERROR REPORTED 17 SLAVES HELD IN VERMONT IN 1790
A clerical error in the office of the United States Census Bureau in its report of the first census taken in Vermont in 1790 makes that report say that there were 17 negro slaves in Vermont that year, as against the generally understood and frequently repeated assertion that no person was ever held in bondage in this state. Vermont declared against slavery in 1777, and that declaration has always been adhered to.
It is true that the printed report of the United States census of 1790 gave sixteen slaves to Vermont, all of them in Bennington County. But it has long been known that that first census, as given to the public, contained numerous errors, and that this assignment of slaves to Vermont was one of them.
The facts are that in consequence of the discovery of many errors in the reports of previous censuses, Gen. Francis A. Walker, superintendent of the census of 1870, instituted a critical comparison of the printed reports of previous censuses with the manuscript returns of the same on file in the census bureau. In the course of this examination Mr. George D. Harrington, chief clerk of the bureau, made the important discovery that in commpiling the returns of Vermont the careless clerk or copyist who did the work transferred the footing of the column of "free colored" persons to the foot of the adjoining column of "slaves." Gen. Walker, in his introduction to ninth census report, noted the discovery in the following words:–
[277]
"A single result of these examinations into the earliest censuses has enough of curious and substantial interest to be noted here. The State of Vermont was, in the publication of the first census, that of 1790, put down as numbering among its inhabitants sixteen slaves. In subsequent publications this number was by a clerical or typographical error changed to seventeen; but with this accidental variation the statement of the first census has passed unchallenged; and antiquarians have even taken pains to explain in what manner it was this small number of slaves should have been found in a State otherwise through all its history a free State. The reexamination of the original census roll of Vermont at the census of 1790, for the purpose of this republicaation, brought to light what had never before been susspected-that these sixteen persons appeared upon the return of the assistant marshal as "Free colored." By a simple error of compilation they were introduced into a column for slaves; and this error has been perpetuated through nearly the whole history of the government until corrected in the accompanying tables." (See page 46 of Introduction to the volume of Population of the Census of 1870.) .
Under the corrected table for Vermont on a subsequent page of that volume will be found the following note: "An examination of the original manuscript returns shows that there were never any slaves in Verrmont. The original error occurred in preparing the result for publication when sixteen persons returned 88 'free colored' were elassified as 'slaves.' "
It is certainly remarkable that this erroneous assignnment of slaves to Vermont should have gone uncorrected for eighty years. It was not because Vermonters of that day did not know better, for the Vermont Gazette, printed at Bennington by Anthony Haswell, in its issue [278] of Sept. 26, 1791, said, "The return of the marshal's assistant for the county of Bennington shows that there are in the county 2503 white males over sixteen years of age, and 2617 under that age; 5559 white females; 17 black males over 4 and under 16; 15 black females. Total of inhabitants 12,254. To the honor of humanity, no slaves."