74 BKITISH LEPIDOPTERA. 



to differentiate in the 2nd stadium, whilst in the 3rd, the tracheae 

 begin to proliferate, and the capillary tracheae or tracheoles arise. 

 The wall of the principal trachea appears to be resolved into fila- 

 ments, and all the secondary branches assume the appearance of bundles 

 of twine. Landoisf regarded them as the product of a transformation of 

 the nuclei, but Gonin thinks they originate from the entire cells, stating 

 that from each cell arises a ball (peloton) of small twisted tubes. As 

 the large branches penetrate into the wing, the balls (pelotons) of fine 

 tracheal threads tend to unroll, and each of the new ramifications of the 

 secondary tracheal system is accompanied in its course by a bundle of 

 capillary tubes. This secondary system of wing- tracheae, then, arises 

 from the mother-trachea at the end of the 3rd stadium, when we find, 

 already formed, the chitinous tunic which will persist through the fourth 

 stage up to pupation. It differs from the tracheoles in not communi- 

 cating with the air-passage ; it possesses no spiral membrane at the 

 origin and takes no part in respiration. Gonin thus sums up the nature 

 of the two tracheal systems in the rudimentary wing, which he calls 

 the provisional and permanent systems, " the first, appearing in the 

 second stage of the larva, comprises all the capillary tubes, and arising 

 from numerous branches passes off from the lateral trunk of the thorax 

 before reaching the wing ; the second is formed a little later by the 

 direct ramification of the principal branch. These two systems are 

 absolutely independent of each other within the wing. Their existence 

 is simultaneous but not conjoint. One is functionally active after the 

 third moult ; the other waits the final transformation before becoming 

 active " (Packard). 



The tracheae (tr., PI. i., fig. 1) penetrate between the two thickened 

 layers of the wing-pad, the outermost layer being destined ultimately to 

 form the upper wall of the future wing, the inner layer becoming the 

 lower wall. The cells which compose the wing-pads are more crowded 

 in the longitudinal direction than in the direction across the wing. The 

 cells forming the tissue at this time are spindle-shaped hypodermal cells 

 (PL i., fig. 2). When the larva changes into a pupa, the wings expand 

 to about sixty times their former area, and the cells, being no longer 

 crowded, lose their spindle shape, and flatten out into a pavement 

 epithelium. 



The wing is evaginated, and becomes an external organ during the 

 quiescent period preceding pupation. Gonin says that in Pieris, six or 

 seven days after the last larval moult, the chitinous wall is formed, the 

 wing remaining transparent. It grows rapidly, and its lower edge 

 extends near the legs. It is now much crumpled on the edge, owing to 

 its rapid growth within the limits of its own segment. Partly from being 

 somewhat retracted, and partly owing to the irregularity of its surface, 

 the wing gradually separates from its envelope, and the cavity of 

 invagination becomes more like a distinct or real space. The outer 

 opening of the alary sac enlarges quite plainly, though without reach- 

 ing the level of the edge of the wing. Forty-eight hours later, exami- 

 nation of a caterpillar which had just spun its girdle showed that the 

 wings had become entirely external, the partition of the sac had dis- 

 appeared, and with it the cavity and the leaf of the envelope. After 

 the destruction of the partition, what remains of the layer of the 

 envelope is destined to make a part of the thoracic wall, and undergoes 

 for this purpose a superficial desquamation. The layer of flattened 



