INTRODUCTION 



II 



the basis of modern systems. In view of the primary importance of the structure of the feet, 

 I have prepared a plate showing the appearance of the tarsi of the front feet in the six 

 principal groups of butterflies. The figures have been carefully drawn from photographs 

 of specially made microscopic preparations, and a description of the special features in each 

 case will be found on the page facing Plate I. 



Regarding the butterflies as of higher development than the moths, Bates utilizes the 

 more or less complete atrophy of the front tarsi as the principal sign of advancement. The 

 first and highest family is the Nymphalidae, which may be described as four-footed butterflies, 

 the front tarsi being only imperfectly developed in both sexes. This family is by far the 

 largest. The second family is the Erycinidae, and in it the first pair of legs are still im- 

 perfect in the male, though complete but reduced in size in the female. The subdivisions 

 of this family have been productive of some difference of opinion amongst systematists. 

 For the sake of simplicity I have in the following table relied only on the palpi as distin- 

 guishing features, and so maintained only two sub-families as adopted by Dr. Sharp and 

 by Professor Aurivillius. The third family comprises the whole of the Lycaenidae. In 

 these butterflies the front legs are only a little smaller than the others, but in the male the 

 tarsus is single- jointed, is covered with spines beneath, and has either no claws or only one. 

 In the fourth, fifth, and sixth families the legs are perfect in both sexes. Many naturalists 

 group the Papilionidae and Pieridae into one family with two sub-families, but the other 

 structural features, notably the absence of the internal nervure in the Pieridae, are so 

 distinct that it seems desirable to separate them more completely. The sixth family, the 

 Hesperidae, are in many features abnormal. In some respects they resemble moths more than 

 butterflies. The bodies are large and the wings small, though the great rapidity of vibration 

 of the latter gives the insects very remarkable powers of flight. The antennae are in most 

 cases not truly knobbed at the end, but terminate in a hook. They are an exceedingly large 

 family, upwards of two thousand species having been described, and they are already 

 regarded by some authors as a separate sub-order under the name Grypocera (Karsch.) 



The following table of classification has been compiled from various sources ^ and forms 

 a summary of the principal recognized features of the various divisions. 



Sub-order RHOPALOCERA. 



A I. Antennae close together at origin ; tibiae of hind pair of legs with a terminal 

 pair of spurs only. 



A. Front pair of legs much reduced in size in both sexes ; the tarsi in the male having 

 only one joint, or sometimes rudimentary joints. In the female having usually five 

 joints, but without claws. Pupa suspended by the tail and hanging freely. 



Family I. NYMPHALIDAE. 



a. Discoidal cell of hind-wing closed. 



b. Fore- wing submedian nervure with a short fork at the base. 



c. Palpi short, slender, and divergent. 



d. Front foot of female ends in a knob. 



e. Larvae smooth with long fleshy tubercles. 



Sub-family i. Danainae. 



1 I am indebted to Trimen's ' South African Butterflies ', ' The Cambridge Natural History &c., for 

 portions of the above and following information on classification. 



B 2 



