( 110 ) 



Hetorocera with rvlindrical (little specialised I) joints we meet witli an irregular 

 transverse row. Moreover, it is often very obvions from the position of the bristles 

 in the Lepidoptera which do not possess a transverse series that these irregularly 

 placed bristles are derived from a more regnlar transverse series, inasmuch as we 

 find very generally, especially on the clnb, that the one bristle stands upon the 

 intersjiace of two others. It is highly interesting to observe that the generalised 

 Hepialid arrangement of bristles obtains, of all Lepidoptera, again in the Lycaenidae: 

 it is certainly not a mere coincidence that just those Iiei)idoi)tera which we had to 

 consider the most generalised in respect to the fine sense-hairs show that belt of 

 bristles all round the joint, while all other Lepidoptera have the dorsal (scaled or not- 

 scaled) and ventral sides differing considerably in the arrangement of these organs. 



The modifications which the regnlar jiostmedian belt of bristles, as found among 

 Tjycaentil'ie. (f 0), nndergoes are of two jirincipal kinds : a reduction in the number 

 of the bristles and an alteration in their position. Both kinds of development may 

 occur at the same time on the same antenna. 



There are many antennae which have the bristles in a generalised state ventrally 

 and specialised dorsally, while the inverse never occurs ; hence it is clear that the 

 specialisation begins dorsally, and that, therefore, the next generalised stage after 

 the Lycaenid-Hepialid antenna is represented by a form in which the ventral snrfiice 

 has the belt of bristles intact, while the bristles are irregnlar in position and reduced 

 in number dorsally. Such a stage of development we meet with largely among 

 Lycnenidae and Ilcspi'riidae (figs. 1, 8). From this second type a third one, fouud 

 among Li/cnenidae, Ilesperiidue, and largely among Kri/chtidne, resulted by an 

 apicad or basad movement of the series of bristles. The apicad movement is com- 

 paratively rare in Lycaenidae (f. T) and Hespcriidae (f. 2), and seems to occur here 

 only in forms which have the bases of the joints depressed, while in Knjcinidae it is 

 the rule, at least as regards the clnb (f Lj). From the fact that certain bristles, for 

 instance the lateral bristles oi Symp/ia/i'dae (figs. 4" — 72), have a diti'erent position 

 in allied forms, while other bristles remain the same in position, we must conclnde 

 that the bristles can change their jilace independently of each other. This changing 

 of position always takes jilace in a longitudinal direction ; a widening and stretch- 

 ing, or a reduction in the length and width of a joint, or a portion of a joint, affects 

 naturally the mutual jrosition of the bristles, but if in allied Lepidoptera the joints 

 are of the same form the bristles move apicad or ventrad, when a change in position 

 occurs. This independent movement of the bristles leads, however, necessarily from 

 a more regular belt to an irregular series, as we find it, for instance, in some 

 Papib'om'dae (f. 39, Leptocircus). In by far the greater number of Lei)idoptcra a 

 reduction in the number of the bristles obtains, and we observe that throughout the 

 whole order it is the most lateral bristle at each side of the ventral surface which is 

 preserved longest (f. .")), and that, if reduction goes on, the interno-lateral one is the 

 first of the two to disappear (f 14). The highest degree of reduction is, of course, 

 reached when all the bristles have obliterated, a stage found largely among 

 Parnaminae. The reduction of the ventral bristles stands in this subfamily 

 perhaps in correlation with the stronger development of (lie dorsal bristles (f. 35). 

 Besides the lateral pair nearly constantly met with in Butterllies and Moths with a 

 reduced series, we find also very often a submesial i)air jireserved. Both the lateral 

 and the submesial pair are often different in position in different groups (^Xi/mphalidae, 

 figs. 47 — 73 ; Fieridw, figs, l.o — 31) ; whereas in the same group their position 

 is often constant, sometimes very constant (compare submesial basal pair in 



