48 
ST EMM A TOIUL US. 
segmented ancestor, and that this apparently anomalous character of 
the Epimorpha is after all, merely a primitive condition which most 
Eabrata have left. Moreover, it no longer throws doubt upon the 
affinity of the Epimorpha with the other Chilopoda, or of these with 
the Hexapoda. 
That a many-segmented embryo was the condition which obtained 
among the remote ancestors of the Labrata may, then, be inferred 
from three considerations : 
1 . A many-segmented condition not being an advantage would 
not be called into existence or favored by natural selection. 
2. This condition appears in two unrelated orders of Labrata 
whose nearer relatives are distinctly anamorphous. 
3. It also appears in the viviparous Malacopoda which are probably 
more nearly related to the ancestors of the Labrata than any other 
extant forms. 
The hexapodous condition of the young of Diplopoda furnishes ac- 
cordingly very slight argument for any direct or genetic relationship 
between the Diplopoda and the Hexapoda, as Bollinan declared, even 
without the support of the present evidence. 
That the embryos and larvae of the ancestors of Stemmatoiulus 
were at one time provided with numerous legs as well as segments is 
indicated by the existence of a nearly complete series of chitinized, 
flattened, lanceolate processes or propodia, more highly developed 
than the so-called ‘ ‘ buds ’ ’ which precede the legs of some other Di- 
plopoda, and doubtless homologous with the ventral appeadages of 
Campodea and other apterous hexapods. We have here additional 
reason to believe what has long been suspected, that these abdominal 
appendages are rudimentary legs. Moreover, the protrusible vesicles 
of Machilis and even the collophore of the Collembola may easily be 
remnants of the likewise protrusible nephridial structures which ap- 
pear on the coxse of all the legs of Colobognatha, and on some of 
those of Coelocheta, but have otherwise disappeared from the Diplo- 
poda. At any rate, these various problematical appendages in Cam- 
podea, Japyx, and Machilis, if ancestral, rudimentary and trans- 
formed from their original function, do not furnish reasons for suppos- 
ing their possessors to have any close affinities, nor break the force of 
the wonderful structural diversity long known to exist between the 
various Thysanuran families. Lubbock pointed out over twenty years 
ago that the association of Campodea and Japyx with the Thysanura 
proper (Lepisma, Machilis) must be considered as merely provisional, 
and subsequent discoveries have rendered their distinctness more and 
more apparent. The proposition of Grassi and Stuinmer-Traunfels 
to place Japyx and Campodea with the Collembola as a suborder En- 
