POLYDESMOIDEA. 111 
to be very material whether family or subfamily rank is given to the groups into which 
the genera are capable of being classified ; nor is it very important that the said groups, 
whatever their systematic rank, should in every case be established upon equivalent 
characters. Where a large number of genera are affiliated by a single character in 
common, it is often convenient to give the assemblage the status of a subfamily or 
family ; and at the same time it is often expedient to withhold such rank, at all events 
temporarily, from a single genus, though it may differ from its nearest allies in 
characters of the same or even greater systematic value than those which in other cases 
have served as the basis for the groups of higher rank. An instance of this will be 
found in the following pages by comparing the Spheriodesmide with the Chelo- 
desmide. In the latter generic value is accorded to structural differences in the 
phallopod which are not of greater taxonomic importance than the structural differences 
presented by these organs in the Spheriodesmide, to which only specific significance is 
given. When the Spheriodesmide are better known it will probably be found that 
Sphweriodesmus may be conveniently broken up into several; but for the present I 
prefer to keep all the species together under the one genus, although the variation 
in the phallopods is much greater than that admitted in any other genus recorded 
in this monograph. 
Complete consistency in the estimate of characters can be gradually introduced with 
progress of knowledge; and our acquaintance with many genera of Polydesmoidea is 
at the present time too incomplete to make finality in the classification of this group 
possible. 
In this monograph I have retained in almost every case the families or subfamilies 
established by my predecessors where I have found these assemblages to be definable 
within the limits of the Diplopod fauna of Central America, leaving alone the further 
question as to their final admissibility until a complete revision of the Polydesmoidea 
of the world has been taken in hand by some competent systematic zoologist with 
sufficient time and material at his disposal for the task. 
The subjoined analytical key to the families here admitted does not represent my 
opinions as to the affinities of these groups. For instance, although the Oniscodesmide 
and Spheriodesmide fall under one heading, a, I believe with Cook that the spherical 
form of the body when rolled, with its concomitant structural variations that they have 
in common, has been independently acquired in the two cases. The Spheriodesmide 
may, as Brdlemann has suggested, be an offshoot of the Chelodesmid group ; whereas 
the relationships of the Oniscodesmide perhaps lie with the groups represented in 
Central America by Lophodesmus and Peridontodesmus. ‘The latter may, perhaps, be 
related to the Chelodesmide, as also are the Strongylosomide and perhaps also the 
Platyrachide, through the genera forming the Kuryurine. 
