102 EMBRYOLOGY OF INSECTS AND MYRIAPODS 



In an allied species, Locusta migratoria, Roonwal (1937) found that 

 the neurilemma probably arises from the outlying ganglion cells them- 

 selves. Baden (1936), however, says that in Melanoplus differentialis 

 "the external neurilemma can be traced to the coelom sacs and therefore 

 . . . seems to be mesodermal." As for an inner neurilemma, Baden 

 questions its existence. Eastham (1930a) states that this tissue in 

 Pieris rapae probably arises from the outermost products of the neuro- 

 blasts. Such neurilemma-forming cells become elongate and insinuate 

 themselves between the epidermis and the nerve cord and also pass above 

 the gangha underneath the yolk. Tiegs and Murray (1938) assert that 

 in the weevil, Calandra oryzae, certain intersegmental median cord cells 

 give rise to the neurilemma. In the honeybee Nelson (1915) admits 

 two possibilities : Either the neurilemma is formed as Heymons suggests 

 for Forficula; or, what is more probable, they are merely transformed 

 ganglion cells. 



Finally, Strindberg (1913) expresses the view that in Eutermes the 

 neurilemma is first formed on the dorsal side of the ganglia and is derived 

 from ganglion cells that unite to form the tissue. Later it is also formed 

 on the ventral side; but since this occurs after the ganglia have separated 

 from the epidermis, it is good evidence that the neurilemma is not 

 formed from the dermatogene layer. The inner neurilemma in this 

 insect develops after the neuropile is wholly covered by the cortical layer 

 of ganglion cells and therefore without doubt owes its origin to the gangUon 

 cells immediately adjacent to the neuropile. Whether or not the cells 

 of the median cord aid in the formation of the outer neurilemma Strind- 

 berg did not determine. The development in Formica and Chrysomela 

 according to Strindberg takes place essentially as in Eutermes. 



The Stomatogastric System. — Heider (1889) was the first to observe 

 that the frontal ganglion and the recurrent nerve originated in an invagi- 

 nation of the dorsal wall of the stomodaeum, an observation later verified 

 by Graber, Carri^re, and Wheeler. Later still, Heymons (1895a) in his 

 studies on the development of the Dermaptera and Orthoptera found 

 that the stomatogastric system in general is derived from three medially 

 lying invaginations in the dorsal wall of the stomodaeum. From these 

 invaginations there arise, first, the frontal ganglion, then an unpaired 

 mass behind the first, which is the common rudiment of the paired ventric- 

 ular ganglia. These three rudiments then become connected with one 

 another by the recurrent nerve. Roonwal's account (1937) of the 

 development of this system in Locusta migratoria agrees in all essential 

 details with that in other Orthoptera and Dermaptera (Heymons, 1895a; 

 Wiesmann, 1926). The only difference is that in all the forms whose 

 development has been studied previously, the ventricular ganglion is 

 unpaired; in Locusta, as in all Saltatoria, it is paired. In Eutermes 



