( 391 ) 



The three carinae are high; the impressions between them vary in depth, width, 

 and ontline. The fine sense-hairs are seldom distributed over the greater portion 

 of the groove ; they form mostly a rounded patch, of often small size (figs. 68 — 70). 



There is a ventro-basal pair of })rist]es, and a median or postmedian lateral 

 jiair. On the dorsal side of the not-scaled joints we observe, besides, another pair 

 of bristles (f 70), which is always subapical or apical on the club. In many 

 species allied to If/iomia, especially in Ilyposcada, the not-scaled joints have a 

 number of additional dorsal and dorso-lateral long bristles. The sides, sometimes 

 the whole dorsal surface (if not scaled), are beset with setiferons punctures. Dorsally 

 the joints are in the mesial line often somewhat impressed: in Ilammln/as there is 

 dorsally an almost complete mesial furrow along the not-scaled jmrtion of the 

 antenna. 



/. DANAINAE. 



All the species of this subfamily have the antennae bare of scales, as has already 

 been observed by Scudder,* who, however, unites the Neotropinae with scaled 

 antennae with the Dnnainae. The three carinae are always developed, but in some 

 forms not so strongly marked on the club (f. 71, D. ple.cippi.is) as in most species. 

 The grooves are on the joints of the stalk basally much more shallow than iu the 

 middle of the joint. The fine sense-hairs are always restricted to a rounded patch, 

 which is very conspicuous (f. 72), especially in species with dark antennae, on 

 account of the silvery hue of the short hairs, and is more impressed than the rest 

 of the groove. 



The ventro-basal and the lateral, subapical or median, pairs of sense-bristles, 

 generally well developed, are in some cases (JJa/iais) obsolete. Dorsally (f. 73) 

 there is regularly a postmedian pair of bristles present, besides one or more bristles 

 which appear occasionally on one or the other joint without regularity. 



From the descriptions of the antennae of the families of Butterflies given iu the 

 foregoing pages, we learn that an antennal organ or structure is variable in one 

 family, while it is relatively constant in other groups. Among the I'api/io/iidae we 

 find species with not-scaled antennae and such with an extended development 

 of antennal scaling, while fn the Ilesperiklae- and in the Dismorpkii/iae the not-scaled 

 area is constantly restricted to the distal joints. The Nymphalidae have a ventro- 

 basal and a lateral pair of bristles in all the members where the bristles are developed, 

 and the Ilespcriulae have always a transverse series of bristles (sometimes reduced 

 to two) ; while the various members of Papilionidai' deviate in this respect widely 

 from each other. The Hesperiidae have never grooves ; the Fierinae have always 

 one, the Dismorphiinae always three, the Nymphalidae two; while the Fapilionidae 

 have one, two, or no grooves. And a similar variability in some and constancy 

 in other groups is found in respect to the fine sense-liairs. It is obvious tliat the 

 families in which a certain antennal organ or structure shows difterent stages of 

 development in the various members will most likely present us also with such 

 stages as will reveal to us, firstly, the line of development by which other 

 families, in whicli that organ or structure is uniform in type, have arrived at their 

 stage of development, and hence siiow us, secondly, the phyletic connection between 

 the various groups of Butterflies. Before, however, entering upon these questions, 



