526 W. L. TOWER, 



short and soon passes into one more characteristic and of longer 

 duration, where the invagination exists as a furrow through the length 

 of the wing disc (PI. 14, Figs. 2 and 7). The occurrence of the pit- 

 like invagination in several species of insects is of interest, and while 

 it may be coincidence, or its existence may be due to similar mechanical 

 causes existing in the wing fundament, it is also possible that it is 

 of wide phylogenetic significance, and this seemed especially so when we 

 consider that the tracheae arise in exactly the same way in the em- 

 bryo, as does this pit in the wing fundament of the larva. At present, 

 it seems to me that this stage is in all probability entirely dependent 

 upon mechanical causes existing in the wing fundaments and is not 

 to be regarded as of phyletic significance. 



The time in larval life when the wings first appear is of interest. 

 In the Lepidoptera as far as is known, the wing discs are found in 

 the youngest larvae and in all probability arise in the embryo. In 

 the Diptera, Weismann and others have found that in some Nemato- 

 cerous forms the wings appear just before pupation, while in Brach y- 

 cerous forms the wings are present in the young larvae and probably 

 arise in the embryo. 



In the Coleoptera there is considerable diversity in the time ot 

 appearance of the wing discs. In H. ISjmnctata the wing appears 

 when the larva is "about one fifth grown" (Comstock & Needham 

 1899) and in a near ally, C. hipunctata^ at the beginning of the third 

 instar. In L. decemlineata, and likewise in E. borealis, the wing 

 appears at the beginning of the third instar. In 0. scdbra careful 

 search failed to show any trace of the wing until the last larval stage, 

 but in a near ally L. fusca et al, the wing is present in larvae of 

 the first year. The Cerambycidae and Buprestidae are also interest- 

 ing as showing a postponement of the beginnings of wing development 

 until the end of larval life, although the wing due may be present, as 

 in 0. femorata, throughout larval life. In the Carahidae the wings 

 appear in the penultimate larval stage, but undergo no development 

 until just before pupation. Lastly, in the Curculionidae the wings 

 appear in the last larval stage. In the following table I have given 

 the time in the larval life when the wings appear and some other 

 data concerning the number of stages for the species studied by my- 

 self and others. 



