2 14 Alexander Petrunkevitch, 



teen species of spiders are new to science and are described here 

 for the first time. Besides the more or less complete specimens 

 the Scudder collection of the Museum of Comparative Zoology 

 contains also four impressions of single legs and eight cocoons. 

 The latter were described by Scudder as belonging to a species 

 which he named Aranca columhiac. I have not seen these speci- 

 mens, nor do I include them in the present study because of the 

 utter uncertainty of classification. That the specimens in ques- 

 tion are spider-cocoons seems to be if not absolutely beyond doubt, 

 at least highly probable. But even the family could not be identi- 

 fied with any degree of certainty, and the figures given by Scud- 

 der in his Tertiary Insects on Plate 2 make one believe that the 

 cocoons in cjuestion are not all of the same species. 



A few words of explanation are necessary concerning the 

 recognition of species. Generally speaking it is much easier to 

 recognize in a new palaeontological specimen a new species than to 

 place it correctly in the system. Our knowledge of American Ter- 

 tiary spiders was based on Scudder's work. A keen observer and 

 a man broadly versed in the systematics of Arthropods, Scudder 

 had no difficulties in assorting his material and describing the 

 new species. He made minor mistakes, in a few cases he placed in 

 the same species specimens distinctly belonging to dift'erent spe- 

 cies, but the bulk of his species stand good to-day. At the same 

 time Scudder has never been a professional arachnologist and the 

 principles of classification gradually arrived at through the work 

 of several distinguished European arachnologists remained 

 unknown to him or did not appeal enough to him to be extended 

 from the field of recent forms to that of extinct ones. Therein 

 lies the weakness of Scudder's work. For although his species are 

 good, the characters upon which their description is based, from 

 the point of view of modern knowledge are entirely inadequate 

 for the purpose of placing the species in the system. As will be 

 pointed oiit further down, not only the generic characters, but 

 even family characters as well as characters of entire sub-orders 

 are usually so obliterated that Scudder had to rely mostly on 

 external similarity. Naturally species of the same family have 

 more resemblance to each other than to species of other families. 

 External resemblance is the first principle in any assortment of 

 large collections. At the same time, as is well known, it is one of 



