24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I07 



PART II. REPRINTS AND REVISIONS 



In 191 1 Sherborn and Durrant stated that the publication of the 

 parts proceeded "with great regularity, commencing January 1824, 

 and finishing December 1839. so the dates of the plates may be ac- 

 cepted with certainty." The implication here and elsewhere that after 

 plate 10 (parts i and 2) four plates were issued each month and dated 

 accordingly, is susceptible to checking only in part by any means at 

 my disposal. However, it did seem worth while to examine all the 

 plates in the first edition to see if the dates on the plates bore out this 

 assumption. This examination revealed no case where the dates do 

 not agree with the monthly schedule. 



This examination of dates was made in the copy in the United 

 States Department of Agriculture library, which is in numerical or- 

 der. In the course of this examination it was discovered that com- 

 mencing with plate 386 many of the plates had no date line at the 

 bottom. These plates were on somewhat different paper from the 

 others and had a different appearance. In all, 89 of these undated 

 plates occurred. 



No explanation being obvious, additional sets were obtained for 

 comparison. The only other set examined which is arranged in nu- 

 merical order ^ also contains many of these undated plates, but they 

 number 103, with only 36 of these on plates corresponding to those 

 in the first set. The remainder of the undated plates are of lower 

 number, starting with 83. It was further noted in this second set 

 that many other plates were obviously redrawn and of poorer 

 workmanship. 



These discoveries led to the examination of two more sets, these 

 both in systematic order.* In neither of these were any of the vmdated 

 plates found, but numerous other inconsistencies added to the puzzle. 

 A tabulation of the five sets in parallel columns in numerical order 

 failed to throw much light on the problem, but certain facts have been 

 made clear. 



1. Unquestionable original examples of all 770 plates have been 

 seen, engraved in high quality by Curtis himself (at least at first) 

 and showing no deviation in style. 



2. Curtis' reprints of plates 1-50 are equally recognizable, the plate 

 number always being underscored (the text was reset and number 

 underscored on 1-34 only). 



3 Through the courtesy of the Rutgers University Library, New Bruns- 

 wick, N. J. 



^ Through the courtesy of the library of the American Museum of Natural 

 History, New York City, and of the Peabody Institute Library, Baltimore, Md. 



I 



