94 Annals Entomological Society of America [Vol. VIII, 
tinct types of respiratory system. In certain families we find 
a series of lateral spiracles (peripneustic type), as in normal 
terrestrial insects. In others the larval tracheal system is 
laterally closed, open spiracles occurring only in single anterior 
and posterior pairs (amphipneustic), or in but a single posterior 
pair (metapneustic). Still other forms show a completely 
closed tracheal system (apneustic), the skin assuming a re- 
spiratory function, sometimes supplemented by specially evolv- 
ed appendages (blood-gills). 
It is evident that the presence of a series of open lateral 
spiracles represents the normal unmodified type and is therefore 
a more primitive condition, while the partial or completely 
closed respiratory system has been acquired in adaptation to 
an aquatic environment. We must therefore look upon the 
dipterous larve with lateral spiracles as an archaic type and any 
classification that disregards this fact must necessarily be faulty. 
From such an ancient type are to be derived the amphipneustic 
and metapneustic forms. As to the forms with wholly closed 
trachee, it must be pointed out that they may be derived equally 
well either from the modified types or directly from the peri- 
pneustic stock, and there is good evidence that such forms have 
originated in both ways. 
While these conclusions have been reached by wholly 
independent reasoning, it must be pointed out that the phylo- 
genetic significance of the modifications of the larval tracheal 
system has been already indicated many years ago. Palmén 
pointed it out in his classic work on the tracheal system of in- 
sects, but he left to others the application of the principle.» He 
argues that the functionless thread-like spiracular branches 
which exist in forms with wholly or partly closed tracheal 
system could not have been always functionless. “It is un- 
thinkable that they were rudimentary from the beginning 
and have originated or been acquired phylogenetically as mere 
threads. On the contrary, they must have been formerly con- 
stituted like the other parts of the tracheal system and have had 
their full function in order to be able to occur now as rudimen- 
tary structures. The larve existing at present which possess 
a closed tracheal system, have therefore formerly had open 
stigmatal branches in all stages of their existence. This means, 
in other words: Out of the relatively primary, open form of 
tracheal system of the oldest insects, there has originated 
