PHYCIODES I., II. 



produced artificially. It was taken at Grimsby, Ontai-io, and probably owes its 

 peculiarities to its chrysalis having been* frozen soon after it formed. That such 

 examples are not frequently seen on the wing may be partly due to the ftict that 

 chrysalids of tender age subjected to severe cold are usually killed by it. Fig. 

 13 is the likeness of an aberration sent me by Professor Garman, and taken at 

 Normal, Illinois. 



Dorfmeister, who experimented on the eiiect of cold upon the chrysalids of 

 butterflies, as related by Dr. Weismann, was led to believe that temperature ex- 

 erts the greatest influence during the turning into chrysalis, but nearly as much 

 shortly after that time. By my experiments it would appear unnecessary that 

 cold should be applied before the chrysalis has fully hardened in order to effect 

 a change of form, and this requires a longer or shorter time in different species, 

 but never more than a few hours. In 1S7G, the chrysalids of Tharos, which 

 were exjDOsed at from six to nine hours after forming, changed as completely as 

 those at from tliree to six houi's. In 1877, all which were exposed at from one 

 to nine hours changed, while some which were exposed at from thirty to sixty 

 minutes were not changed at all, though others were. The turning-point is evi- 

 dently not at or immediately after the forming of the chrysalis. The cold com- 

 pletely retards the development of the butterfly, and at least the full natural 

 chrysalis period passes after removal from ice before the butterfly emerges. This 

 has happened in eveiy case under observation, even when from insufficient or 

 intermittent cold no change of form has resulted. 



The change of form seems to be almost wholly a matter of coloration, for the 

 shape of the wings is not altered. The natural shape of Weds Jill is quite difler- 

 ent from that of MarceUus, but when the latter by application of cold is forced 

 to assume the markings of Walshil, it retains its own shape nevertheless. In 

 the other species experimented on there is no diflerence in shape between the 

 winter and summer forms. The coloring process in all chrysalids does not set in 

 till just before the butterfly emerges, no matter how protracted the pupal period. 

 It may be days or weeks, and in the case of over-wintering chrysalids, it is many 

 months. On removing one of the wing cases, soon after the chrysalis is formed, 

 the wing is seen to be a delicate, transparent, colorless membrane, ci'ossed by 

 white veins. Very gradually this membrane thickens, losing its transparency, 

 and becomes of a dull white. A few hours, or one or two days, before the close 

 of the period, it grows darker, then a faint discoloration comes on, and the pat- 

 tern begins to show itself, the colors becoming strong and the markings well- 

 defined as the disclosure is about to take place. On removing the wing case 

 from a chrysalis just oft' ice, the wing presents itself in its colorless and trans- 

 parent stage ; that is, the development has been arrested by the cold, and the 



