39 



beyond the region of safety ; so that we may say, speaking generally, 

 the further the larva comes out of its case the more uncomfortable it 

 feels, and the more certainly it will quickly retire. 



These actions, which I have ventured to interpret on psychological 

 grounds, may be easily observed with a magnifying glass by watching 

 the insects at their work. It is not well to rely upon general observations 

 such as these which may easily have a too great significance put upon 

 them, seeing how small the insect is, and how quickly all its actions 

 are carried out. To prove that the bifurcating hairs are responsible 

 for a series of actions so complicated and have such a high psychological 

 significance, it is but natural that more tangible evidence should be 

 demanded. Fortunately the arms which this insect builds enable the 

 truth of these observations to be tested in a very definite way by 

 measurement which cannot well be disputed. 



In order to build the arms the larva must come out as far as the Proof by measure- 

 arms extend ; hence the length of the arms is a sure indication of the 

 amount of the body that is willingly extended outside the case. The 

 actual measurements in four different instances are as follows : 



Measurements. Average, 



mm. mm. mm. mm. mm. 



Length of larva ... ... 4*6 4-2 4^5 4'! 4*35 



Length of head and nine 



segments... ... 37 2'6 3-8 3-1 3-3 



Length of arm ... ... 2'g 2*4 2-5 2'5 2*57 



Difference between length 

 of head and nine seg- 

 ments and length of arm *8 -2 i'3 7 7 



These figures show that not only is the arm much shorter than 

 the length of the animal but that it is also decidedly shorter than the 

 head and first nine segments, the average difference being 7 mm. 

 In other words the larva can reach the extremity of its arms without 

 exposing all the first nine segments, that is without the last pair of 

 bifurcators losing contact with the case. 



