132 Mr. P. A. Buxton on the 



I believe that the characters by which the three roots 

 may be separated are these : The inner root {in. r. ; inner e 

 Wurzel, Kiihnle; tubercule interne, Dujardin; Balken, 

 Flogel) runs backwards and inwards and terminates 

 between the middle lobe and the inner capsule of the 

 central body. Its end is adjacent to that of its fellow on 

 the opposite side. It appears that this root is found in 

 nearly all insects. The forward root {as. hr. ; vordere Wurzel, 

 Jonescu; Vorderhorn, Flogel; anterior root, Kenyon; 

 tubercide anterieur, Dujardin) runs forwards and upwards 

 and ends " free " on the surface of the protocerebral lobes, 

 either under the ganglion cells or else directly beneath the 

 neurilemma. The backward root {po. hr. ; riicklaufige 

 Wurzel, Kiihnle) runs backwards and terminates in the 

 posterior part of the protocerebrum without ever reaching 

 the surface. In many insects either the forward or the 

 backward root is absent; this has led Kiihnle to believe 

 that they are identical. 



I should like to take this early opportunity of answering 

 one objection which will probably be made to the hypothesis 

 that the mushroom body in its typical development pos- 

 sesses three roots. It is well known from Kenyon's work 

 by the Golgi method on the brain of the bee, that the axons 

 which compose the stem branch dichotomously, and that 

 the two branches form the two roots of the mushroom body 

 of that insect. Now it may be urged that this division of 

 the axon into two, which is probably characteristic of the 

 nerve cells of the Arthropoda in general, would find its 

 outward expression in a mushroom body with two roots. 

 To this I may, however, reply that there is no difficulty in 

 supposing that each fibre as it divides supplies two of the 

 three roots ; and at any rate the difficulty remains whether 

 the ascending and posterior branches be one root or two, 

 for the plain facts of their anatomy can hardly be disputed. 

 Furthermore, we are already famihar with the division of 

 the roots in the brains of other insects ; now if the fibres 

 which compose a root can be grouped in such a way as to 

 produce a bifurcation of the root, why should not the fibres 

 of a stem be so grouped as to supply almost any number of 

 roots ? Moreover, Kiihnle has already described the mush- 

 room body of a Phasmid, and of a Termite, both of which 

 had three roots, though he failed to grasp the bearing of 

 this fact upon the general question of homology. The fact 

 of the existence of three roots to the nmshroom bodv is 



