OF WASHINGTON, VOLUME XVIII, 1916 167 



and that of spinning the cocoon. We think, therefore, that this 

 series of molts, whenever it occurs, should be considered as a dis- 

 tinct hypermetamorphic stage. In other words the flat sapfeed- 

 ing type should be accorded the same rank of distinction as that 

 attributed, in the Meloidae for instance, to the campodeiform 

 larva. The cases are, of course, not strictly parallel. The tri- 

 ungulin owes its present existence to the preservation of a primi- 

 tive form, whilst the sapf ceding larval type is the effect of the 

 preponderance of a new form developed at the expense of a pre- 

 existing type. But the preservation of the primitive form in 

 Meolidae is due to biological factors of the same nature as those 

 tending to establish the new modification in Gracilariidae. These 

 assertions are based especially on recent studies of the transfor- 

 mations of the genus Marmara. Between the feeding and spin- 

 ning stages of this genus there is a period of quiescence. This 

 period, it was found, is marked by a special stage similar to the 

 pseudo-pupal stage of the Meloidae. This intermediate stage is 

 coarctate throughout its existence. Shortly before the complete 

 transformation of the spinning larva, the larval heads of the 

 pseudo-pupal and spinning stages may be seen enclosed within 

 the skin of the last feeding stage. Upon emergence, the last larva 

 casts the skins of both preceding stages at the same time. The 

 pseudo-pupa bears the same relation to the prepupal larva as the 

 pupa bears to the imago. It is, essentially, a stage of disintegra- 

 tion and reconstruction of tissues incidental to the profound alter- 

 ations in the organs of the larva. All appendages are merely 

 outlined in the external pellicle, nevertheless, the parts are dis- 

 tinct and are easily homologized. The characters of the feeding 

 type are lost, those of the spinning type are foreshadowed, espe- 

 cially in the labrum and labium. The maxillae and mandibles ex- 

 hibit in their general appearance a more distinct return to the 

 form of the primitive Gracilariid larva. This applies also, in a 

 measure, to the general outline of the head-capsule. Only few 

 data concerning the duration of the pseudo-pupal stage have 

 been obtained. One specimen of Marmara fulgidella was observed 

 on April 19 and 20 of this year. Feeding stopped on April 19 

 about 3 p.m. The larva remained motionless for several hours. 

 The heart-action was regular and occasionally there was a slight 

 jerking of the mandibles. About 8 p.m. the heart-action was 

 considerably slower and the larva was opening and closing its 

 mandibles vigorously. At 12 p.m. the heart-action had stopped 

 almost completely and the larva appeared to be dying. On 

 April 20 at 8 p.m. the pseudopupa was well dovelopecl, th:? only 

 sign of life was a slow pulsation in the region of the ninth abdom- 

 inal segment. The larva remained in the same condition until 11 



