428 ALEXANDER PETRUNKEVITCH. 



tinction and spiders were divided into Terricolae, Tubicolae, Ci- 

 tigradae, Saltigradae, Laterigradae, Retithelse, and Orbithelae. It 

 was soon recognized, however, that orbwebs are constructed by 

 some spiders anatomically very different from Argiopidae and 

 since separated from the latter into a family Uloboridas among 

 the supposedly more primitive division of Cribellatae. The same 

 was done for the Dictynidae among the Cribellatae, which were 

 originally placed with the Retithelae. The similarity in web con- 

 struction was explained by the assumption of parallelism in 

 both groups. Similar though not such pronounced parallelism 

 was found in the case of various Tarantulae or Theraphosid 

 spiders, the habits of which are comparable with those in true 

 spiders. Parallelism of such kind may be easily explained as 

 adaptive to similar conditions of life. But it would be difficult 

 to conceive the construction of such a complicated and perfect 

 snare as an orbweb as a manifestation of habits acquired inde- 

 pendently by every genus of Argiopid spiders. The general fea- 

 tures of this habit must have been acquired before the splitting 

 up of the original stock into new species and persist notwith- 

 standing the fact that the family spread all over the world from 

 the tropical rain forests to the cold zones of the North and South, 

 and comprises an enormous number of species as unlike each 

 other in size and appearance as Nephila with its inch and a half 

 long body and four-inch spread of legs and Theridiosoma scarcely 

 a tenth of an inch in length, Gasteracantha with her abdomen 

 much wider than long and adorned in some species with long, 

 curved spines exceeding many times the length of the body and 

 Tetragnatha with her soft abdomen often many times longer than 

 wide and legs so drawn out and kept in a peculiar fashion that 

 the spider in its web has the appearance of a little twig. All 

 these modifications of structure must have therefore appeared 

 later and influenced only the subordinate features of the habit, 

 the details of the general plan of structure of the geometric web. 

 It was this among other considerations that inclined me to the 

 view which I still hold that the Argiopidae are descendants of 

 Uloboridae. 



The manner in which the female takes care of the cocoon with 

 eggs is another habit characteristic of whole families. In fact so 



