and on the Number of Abdominal Segments, in Insects. 175 



to Prof. Huxley, are situated before the mouth ? We get no in- 

 formation from the author, as the terga of these segments are 

 not traced by him. 



But leaving the point whether the eyes and antennas are ter- 

 gal or sternal appendages in Aphis, sufficient proof can be given 

 that it is inaccurate to state that a pair of appendages is always 

 dependent upon a separate segment. It is a general law, that 

 insects leave the egg with the full number of their segments, 

 and that no segment is ever added during the metamorphosis, 

 though some of the abdominal ones may disappear externally in 

 the perfect state. A great number of larvae, however, leave the 

 egg completely blind, and even destitute of antennas, with a head 

 on which no trace whatever of a division into subsegments can 

 be discovered. How, in this case, can the subsequently deve- 

 loped eyes (which are here certainly tergal, and not sternal ap- 

 pendages) be considered as appendages of a proper segment 

 which has never existed ? And how are the ocelli to be accounted 

 for, which, like the wings, make their appearance first in the 

 perfect state, but which cannot, like the wings, be referred to a 

 proper pre-existing segment ? 



The abdominal segments of Julidce, provided with a double 

 pair of legs, furnish us with another proof that the same seg- 

 ment may bear more than one pair even of ventral appendages. 

 To dispose of this proof it is usually asserted that there are two 

 segments united into one; but not a single observation on the 

 production of these segments during the growth of the animal 

 has been made which supports that theory. On the contrary, 

 I conclude, from Newport's observations and figures, that the 

 new segments which are added to the abdomen appear always 

 from the first with two pairs of legs, and never show a division 

 into two. 



If, therefore, the number of appendages is not a just test of 

 the number of segments entering into a part, we have to inquire 

 what are the requisites of a segment, when we undertake to settle 

 the question whether the head is made up of one or of several 

 segments. It will meet with no opposition when, in the first 

 instance, we require that a segment must be marked by a trans- 

 verse line of demarcation on the integument of the animal, at 

 least in an early stage of its development. Another requisite of 

 a segment is, that it must constitute a ring composed of a dorsal 

 and a ventral arcus. A segment requires, further, normally, a 

 set of proper muscles and a ganglion. A ganglion need not be 

 traced in each segment of the perfect insect, where a number of 

 segments are united to a greater section of the body (thorax or 

 abdomen) ; but in the larva, with homonomous segments, the 

 nervous system forms regularly a chain of ganglia corresponding 



