2S8 



THE AMERICAN ENTOMOLOGIST. 



live importance, but the spinneret is well 

 developed, as also are the maxillae, and 

 both pairs of palpi. If now we pass on to 

 the pupa and imago, we find that the man- 

 dibles, labrum and labium have disap- 

 peared, or are greatly aborted and func- 

 tionless, whilst the maxillae and one or 

 both pairs of the palpi are relatively 

 greatly developed. In other words, pre- 

 cisely those organs which have the greatest 

 development in the earliest larval stages 

 are aborted and functionless in the imago, 

 and vice versa j whilst the second or ordi- 

 nary form of trophi is an intermediate state 

 in which all of the organs are relatively 

 well developed. Reverting to the six genera 

 before mentioned, in which alone, and only 

 in their earlier stages, the first form of 

 trophi has as yet been found, we find that 

 in Phyllocnistis the larva passes through 

 only two stages before becoming a pupa ; 

 the first is its active feeding stage, lasts 

 about three days, and is characterized by 

 the first form of trophi (Fig. 127,) and by 

 the absence of all true legs, and by the 

 presence of a retractile membraneous ap- 

 pendage, or pseudopodium, slanting down- 

 wards from each side of each ventral 

 segment, except the last, from the apex of 

 which there are two long ones. In its second 

 stage, which lasts not more than a day, it 

 has trophi of the second form, but with all 

 of the organs except the spinneret aborted 

 and functionless (Fig. 128 a). It still has 

 no true legs, only the lateral and anal pseu- 

 dopodia above mentioned. It pupates in its 

 mine, and the pupa has maxillce and labial 

 palpi, jointed legs and wings. Fig. 128, b, 

 represents the head of an immature pupa. 

 The organs (/) antennae, ((/) maxillae and 

 (e) labial palpi, filled the empty cast {a) of 

 the larval head, terminating in the corres- 

 ponding parts thereof ; but in mounting 



the floor of the mouth. This is incorrect; there are no max- 

 illae, and what I then supposed to be the maxillje was the 

 toothed and ciliated labrum, as I have since demonstrated by 

 other dissections. Organs of the dimensions above given are 

 dithcult to dissect with instruments, as they break and tear, 

 and their true relations may be misunderstood. But if the 

 head of the larva is cut off (or its cast-off skin will do), and 

 mounted as for microscopic examination in glycerine and gum 

 water (Farrant's medium of the microscopists), and set away 

 for a few hours until it is permeated by the fluid, which has 

 had time also to become stiff; and if the slide and glass are 

 then clasped tightly between the thumb and finger, and a turn 

 given to the sliUe, the mouth-parts will separate so that each 

 will be shown separately and distinctly. 



the specimen for preservation as a micro- 

 scopic object, they have been pressed back 

 and thrown out of position. (<:) represents 

 the cephalotheca. In the first form (Fig. 

 127) the labium is large and capable of 

 great protrusion and eversion, and they are 

 presented in this condition in the figure ; 

 when they are retracted the rod, a, disap- 

 pears, being concealed in a groove in the 

 side of the rod b. 



The labium of Gracilaria robiniella in its 

 first larval stage resembles that of Phylloc- 

 nistis, and, as just stated, all the trophi oiCor- 

 iscium resemble those oi Phyllocnistis in that 

 stage. The labrum of G/-(?(77ar/Vr, Fig. 131, 

 is, however, quite different from that of 

 Phyllocnistis, Fig. 132, and Coriscium, and 

 resembles that of Ornix, Fig. 130. But the 

 labium of Ornix, as shown in Fig. 143, is 

 very different from all the others. In all 

 their stages, except the first larval one, the 

 larvae of Gracilaria, Coriscium and Ornix 

 are very closely allied, and differ widely 

 from Phyllocnistis, resembling rather the 

 cylindrical larva of Lithocolletis. 



The next genus is Lithocolletis, on which 

 I have already commented. Fig. 126 rep- 

 resents the first form of trophi in this 

 genus ; in the first three stages of the 

 cylindrical group, and in the first five of 

 the other two groups ; Fig. 124 represents 

 the second form in the last four stages of 

 the cylindrical group, and in the last two of 

 the (?;■««/<■//« group; whilst Fig. 125 repre- 

 sents it in the last two stages of the flat 

 group. The changes which take place in 

 the genus Leucanthiza are identical with 

 those in the ornatella group of Lithocolletis. 

 The genus is nearly allied to Lithocolletis, 

 but shows some structural resemblance to 

 Phyllocnistis. I was formerly led to the 

 conclusion that the changes in the trophi 

 occurred in the genus Gracilaria at the 

 third molt ; this, however, is an error ; in 

 all of the species that I have examined in 

 this genus, and in the very closely allied 

 genera Coriscium and Ornix, it takes place 

 at the first molt. The larvae in the first 

 stage have trophi of the first form, and in 

 the second and subsequent stages of sec- 

 ond form. I have not seen the larva of 



