Vol. xxii] ENTOMOLOGICAL NEWS. 453 



Fig. 7) ; the origin of the mid-dorsal rectal trachea was not deter- 

 mined. The three non-cellular longitudinal bands of the rectum had, 

 when the viscus was first opened, the shiny, silvery appearance which 

 suggests enclosed air, but no tracheae could be followed into these 

 bands. The chitinous lining of these bands was much wrinkled into 

 short folds. 



The crop of this larva (No. i) contained fragments of a smaller 

 larva of the same species which belong undoubtedly to the small larva 

 referred to on page 407 as having been collected at the same time; 

 these fragments were identifiable as follows: A large part of the head, 

 most of the mentum with left labial lobe and much of the right labial 

 lobe still attached, groups of two or more abdominal segments, and 

 two of the caudal gills. The crop also contained a mosquito larva 4.2 

 mm. long, comprising head, thorax and abdomen all in one piece ; 

 and most of the body of a copepod measuring .3 mm. long. The size 

 of these fragments is such as to excite wonder that they were not 

 more minutely subdivided by the mouth-parts, unless indeed their size 

 is due to unusual 'bolting' by the larva under the stimulation of cap- 

 ture. Behind the gizzard the contents of the alimentary canal were 

 much smaller particles and if the food fragments in the crop were 

 not abnormally large, there is here more evidence of a subdividing 

 function in the gizzard than I was able to adduce on another occa- 

 sion (Calvert, 1899, p. 416). 



The arrangement of the principal tracheae is indicated in PI. XVII, 

 Figs. 7, 8. It does not differ in any important features from that 

 shown in a large-scale (unpublished) drawing which I made many 

 years ago from a living larva, 2 mm. long, of the European Agrion 

 puella. The chief tracheae of two caudal gills have also been figured 

 (PI. XVII, Figs. 12, 13) from the preserved material. Unfortunately, 

 no study of the finer tracheoles was made from life, when alone they 

 can be distinctly traced. 



The ventral nerve cord consists of suboesophageal, pro-, meso-, 

 and metathoracic pairs of ganglia (these last three in their respective 

 somites and very distinct from each other), and seven pairs of ab- 

 dominal ganglia as follows: the first in the middle of segment I, the 

 second near the hind end of 2, the third at the articulation of 3 and 

 4, the fourth to the seventh in the extreme anterior ends of 5, 6. 7 

 and 8 respectively. 



Developmental Changes in the Larva. 



Comparison of larvae Nos. 48, 49, 54, 3 and that (No. 101 ) 

 recovered from the crop of No. I and the exuviae of 58 and 

 6 1 listed on page 449, have furnished only the following differ- 

 ences from the above description of the mature larva : 



