natural; and if it were broken up by attaching undue importance to very subordinate characters, no partial change would suffice; for any principle that might be thought to justify the establishment of one genus, would render it necessary, if consistently acted upon, to create not fewer than thirty or forty. One of the most obvious differences among the species is the presence or absence of a tail; but an arrangement founded on this circumstance, separates, by a wide interval, kinds which are in other respects most closely allied. Nay, the tail itself is often more or less developed in the same species, being sometimes distinct in the one sex while it is inconspicuous or wanting in the other; its value as a diagnosis of genera is thus in a great measure destroyed.
Considerable differences likewise prevail in the appearance of the caterpillars, but these are too imperfectly known to be made the groundwork of an arrangement, even if they were likely to be available for such a purpose by indicating natural groups or affording additional means of characterising them. "Some of them," says Dr. Boisduval, "such as those of Machaon, Alexanor, Asterias, are cylindrical and smooth; others (Crassus, Philenor), are protected with rather long fleshy prominences; in a very great number (Pammon, Memnon, Chalchas, &c.) the two first segments are attenuated, and capable of being retracted under the third and fourth, which are dilated and often ornamented with ocular spots analogous to those presented by
