the larvae at this time. The catching of prey was studied in the laboratory; 

 the larvae were placed in a glass tube of about 3.5 cm width; some small 

 stones were placed in the tube and sand was cemented to the wall of the 

 tube; a rubber tube ending in a funnel was placed in the tube; and it was 

 thus possible to create a water current in the tube; the current could be 

 regulated and animal food was added. When the current was interrupted, 

 the larvae did not notice C yc lop s and Daphnia near them; however, 

 when the current began, the larvae caught the Crustacea. The small larva 

 caught large Daphnia almost half as long as itself; the larva also caught 

 Cyclops and hydrachnids with a strong carapace. During the time the 

 larva catches prey, it collects also detritus and plankton, keeping it on its 

 femoral combs, and eats algae, zooplankton and small plant fragments. 



Growth, molting and number of larval instars 



The molt of the caddis fly larvae, especially that of Annulipalpia has 

 hardly been studied. The larvae of Rhyacophila were sometimes found 

 hidden among loosely connected sand grains or small stones, quite different 

 from their massive pupal case. These primitive structures are probably 

 temporary; they are attached to the substrate, built by larvae living in 

 flowing water for molting. 



The molt of the case-building larvae of Annulipalpia and of all Integri- 

 palpia takes place inside the case. The larva attaches the anterior margin 

 of the case to an object; it then shuts itself in the case with a thin membrane 

 at the anterior opening; some species cover the anterior end with a special 

 hood of sand grains (Silo nigricornis). Species living in calm 

 65 habitats, e. g., C r u no e c i a i r r o r at a (at the edge of springs, among moss 

 or humid fallen leaves) do not attach the cases for molting, as there is no 

 danger of being carried away. 



During the relatively long larval stage of caddis flies, the larva has 

 4 molts, so that there are 5 stages. This number of stages is characteristic 

 for the order; the larva of some species, e. g., A g apet u s fuscipes, 

 molts 6 times, so that there are 7 larval stages; the larvae of Sericosto- 

 ma molt 5 times, i. e., 6 larval stages (Nielsen, 1942:262-263). The 

 penultimate molt (into the pupa) takes place in the case; the last molt 

 (into the imago) takes place when the insect passes from water to land. 



The growth rate of parts of the body is not uniform. The soft abdomen 

 of the larva grows continuously during all stages; on the other hand, 

 sclerotized parts of the body (head, prothorax, legs and anal legs) grow only 

 when the larva molts, during the interval in which the old integument has 

 been shed and the new one has not yet hardened. The length of the body 

 cannot be used as a character of the stages, because the abdomen increases 

 in size during each stage, while the size of the sclerotized parts, especially 

 the width of the head, remains constant. The width of the head is a reliable 

 morphological index to distinguish stages of larvae in the population. 



The body of the larva is more or less transparent during the whole 

 1st stage; head, thorax and legs are of a relatively large size, but the 

 abdomen is short and narrow; the 1st abdominal segment is without 

 tubercles in Integripalpia. The sclerites are thin, the sutures are absent 



59 



