296 SIR JOHN LUBBOCK ON THYSANURA. 
there are no other special organs for respiration, and the process seems to be carried 
on through the skin. Moreover, while Papirius, like most of its allies, has no trachez, 
Smynthurus has a well-developed system. I know hardly any other case of species, 
so closely resembling one another in other respects, differing so entirely as to one of the 
most important parts of their internal anatomy. 
The structure of the mouth, as far as it goes, is unfavourable to the view of those who 
regard the Collembola as true insects. 1 quite agree with Dr. Meinert that the mouth 
differs essentially from both the principal types found among insects, without, however, 
making any near approach to that of the M yriopoda or the Arachnida *. 
The presence of a saltatory caudal appendage must be taken for what it is worth. It 
is very remarkable that no similar apparatus is possessed by any one of the almost iunu- 
merable insects, many of which, however, possess the power of leaping in a high degree. 
Nor, on the other hand, do any of the Collembola jump like Gryllus, Haltica, or Pulex, 
by means of their hind legs. The true value of such a character as this, however, is as 
difficult to estimate as it it is easy to apply. à 
The same observations apply to the gastric tube, which, as I have already observed, is 
even more characteristic of the Collembola than the caudal appendage. 
As the upshot of allthis, then, while the Collembola are clearly more nearly allied 
to the Insecta than to the Crustacea or Arachnida, we cannot, I think, regard them as 
Orthoptera or Neuroptera, or even as true insects. That is to say, the Coleoptera, Or- 
thoptera, Neuroptera, Lepidoptera, &c. are, in my opinion, more nearly allied to one 
another than they are to the Poduride or Smynthuride. On the other hand we 
certainly cannot regard the Collembola as a group equivalent in value to the Insecta. | 
If, then, we attempt to map out the Articulata, we must, 1 think, regard the Crustacea 
and Insecta as continents, the Myriopoda and Collembola as islands—of less importance, 
but still detached. 
Or if we represent the divisions of the Articulata like the branching of a tree, we 
faust picture the Collembola as a separate branch, though a small one, and much more 
closely connected with the Insecta than with the Crustacea or the Arachnida. 
The Collembola fall into five well-marked families. "Two of these (the ada 
and Poduride) have long been recognized. The former, however, must certainly be 
divided into two, the Smynthuride and Papiriidee, characterized respectively by the pre- 
sence and absence of tracheæ. It is true that the former family wil contain only two 
genera, the latter only one; but if our classification is to have any scientific value, it is 
impossible to place in one family species which differ so much in their internal anatomy. 
The Poduridæ must be divided into three families—the Poduridæ proper, the Lipuridæ, 
and the Anuride. The two former have mandibulated mouths, the latter suctorial; 
the former is saltatorial the two latter do not possess the characteristic spring. The 
* Isay any “near ” approach, because as Mr. Humbert has pointed out to me, the mouth parts of Scolopendrella 
in some points approach those of the Collembola. This genus, indeed, appears to be more interesting and peculiar 
than either Gervais or Newport supposed. For instance, it has on the underside of each segment a pair of appen- 
dages closely resembling those of the Lepismidæ—a fact which ae doubts whether the subabdominal appendages 
of that en. represent the legs of Myriopoda. 
