existence of these new natural suborders also in the larvae and described 

 them; however, he did this "by the way," not considering it an important 

 taxonomic problem. This was repeated also in his latest work (Siltala, 

 1907:596). He understood that the suborders of Kolenati were artificial; 

 he did not use them, but used his own views in the systematics of the order 

 which he based on the biology of the larvae. This induced him, in 1906—1907, 

 to introduce a division into suborders which was later established more 

 definitely by Martynov (1924:18—20). Siltala thus anticipated the revolution 

 in the system of caddis flies by Martynov (1924). This revision consisted 

 of a division into natural suborders instead of the artificial division. 



In the second of his "Untersuchungen, " Siltala criticizes the artificial 

 135 suborders of Kolenati, stating decisively on the family Leptoceridae: 

 "Die Leptoceriden, die mit den kocherlosen Formen, als Aequipalpia 

 vereinigt werden, haben mit diesen nichts zu tun" (Siltala, 1907:606). The 

 foregoing issue of the "Trichopterologische Untersuchungen" is a large work on 

 the postembryonal development of caddis fly larvae, forming a supplement to the 

 "Zoologische Jahrbucher" (Siltala, 1907). The major part of this work deals with 

 the number of molts and the molting process of caddis flies, duration of the larval 

 stages, respiration, feeding and movements of the 1 st-stage larvae, construction 

 of the case by the young larva; it also describes the cutaneous appendages of the 

 larvae, divided by the author into "series with setae" and "series with spines"; 

 he also describes the primary and secondary chaetotaxy. 



This work deals with the postembryonal ontogenesis in a large group of 

 families and subfamilies; it describes the characteristics and specialization 

 of the primary chaetotaxy and the development of the secondary chaetotaxy 

 in the different larval stages. Siltala gives a comparative -morphological 

 survey of a number of larval structures, including gula, antennae, mouthparts, 

 anal legs and primary and secondary chaetotaxy. 



The author lists the families and subfamilies proposed by Ulmer (1906), 

 but trying to find a natural order in the system, changes their sequence and 

 places the group of families with campodeiform larvae before the group of 

 families with eruciform larvae because he considers the former as more 

 primitive. He arranges the families in each group from the most primitive 

 to the most specialized; he places the families Rhyacophilidae to the Hydro- 

 psychidae in the group with campodeiform larvae; and the families 

 Phryganidae to the Sericostomatidae in the group with eruciform larvae. 



The marked resemblance between the 1st -stage larvae of the different 

 families and subfamilies (especially the constant number and arrangement of 

 the primary setae in the different groups) is considered an indication of 

 their common ancestry by Siltala, and an expression of the unity of the 

 group as a well-defined branch of the class Insecta. He proposed to 

 determine the degree of primitivity or specialization of the various groups 

 by the degree of resemblance or lack of resemblance between the full-grown 

 larva and the young 1st -stage larva. 



The author places the group of Trichoptera with campodeiform larvae 

 (the future Annulipalpia Mart.) below the group with eruciform larvae (the 

 future Integripalpia Mart.). 



After examination of this rich material and study of the early stages of 

 ontogenesis (which is very valuable for the solution of problems of 

 phylogeny), Siltala discusses the phylogenies of a number of families and 

 subfamilies; he compares the morphological characters of the specialized 

 full-grown larva and the primitive 1 st-stage larva. 



127 



