242 [October, 



the vaguest hypothesis as to why the inemliers, say of the g'enus 

 Pfii/lliodes, should he monotonously uniform, and those of Lomjttarsus 

 so wildly irregular. 



It will be plain then that, since our knowledge of the members of 

 this group is virtually confined to their imaginal form, and that only in 

 a morphological sense, discrimination of species becomes somewhat 

 empirical, or even to a large extent provisional. 



In the notes, therefore, which follow, it is hoped that this con- 

 sideration will be understood as implicit. Throughout them an attitude 

 possibly more synthetical than that of some recent Continental students 

 of the genus has been adopted, it Ijeing the belief of the authors that 

 errors of defect invariably more easily admit of rectification than do 

 those of excess, and that nothing is more easy or more futile than to 

 apply a fallacious analytical method to such a group as this. 



We have devoted considerable time to a study of the genus, and 

 our investigations have been much facilitated by the very kind assist- 

 ance of friends and fellow workers (to be mentioned more particularly 

 later on) , who have placed their observations at our disposal, communi- 

 cated specimens, and allowed the examination of collections. We may 

 here, perhaps, specially allude to a box containing exponents of the 

 genus, kindly communicated to us by Mr. P. de la Garde, discovered by 

 him among duplicate boxes left by the late Mr. T. V. WoUaston, which 

 internal evidence clearly shows to have been sent to that gentleman by 

 M. Allard in 1861. As the original list of names of the contents was 

 found with, and accompanied the box, we were enabled to verify and 

 compare with others what are practically Allardian co-types of the 

 genus, and this we need hardly say we have found of invaluable assist- 

 ance in our investigations. 



The problem of food plants, a knowledge of which would 

 seem to be of the utmost value as providing physiological data in the 

 appreciation of species, has not been overlooked. The difficulty, how- 

 ever, in associating, without the possibility of error, forms so agile and 

 elusive with any special plant is formidable and exacting. We have, 

 however, obtained some little fairly assured knowledge on this subject, 

 and its patent incompleteness will, we hope, stimulate other students 

 of the group to attempt the solution of the problems which still remain. 



Before proceeding further it may not be out of place to say a word 

 about the characters which have been generally used in this genus for 

 the differentiation of its species. 



Colour. — Within rather wide limits the Longitarsi are fairly con- 



