304 MOLLUSCOIDEA SUB-KINGDOM V 
While his classification lacks a perfect understanding of the features in question, 
it is remarkable that von Buch, nearly sixty years ago, and Deslongchamps, 
twenty-eight years later, recognised some of the principles upon which the 
classification of the Brachiopoda is now established, viz. the nature of the 
pedicle opening. 
Up to 1846 the general external characters of the Brachiopods served the 
majority of authors as the essential basis for generic differentiation. In that 
year, however, King pointed out that more fundamental and constant 
characters exist in the interior of the shell, a fact which soon came to be 
generally recognised, mainly through the voluminous and admirable contribu- 
tions of ‘Thomas Davidson. 
Waagen in 1883 found it “absolutely necessary” to divide Onan two 
orders into seven sub-orders. The basis for these sub-orders rests on no 
underlying principle of general application, and yet five of these divisions are 
of permanent value, for each contains an assemblage of characters not common 
to the others. 
No classification can be natural and permanent unless based on 1 the history 
of the class (chronogenesis) and the ontogeny of the individual. However, as 
long as the structure of the early Palaeozoic genera remained practically un- 
known, and the ontogeny wholly unrevealed, nothing of a permanent nature 
could be attempted. In the recent and very excellent volumes by Hall and 
Clarke (Palaeontology of New York, vol. viii.), the great majority of the 
Palaeozoic genera are clearly defined. The ontogenetic study of the Palaeozoic 
species was initiated in 1891 by Beecher and Clarke, followed by Beecher, 
and more recently by Schuchert ; and their results combined with those derived 
from the study of the development of some living species, such as have been 
published by Kovalevski, Morse, Shipley, Brooks, Oehlert, Beecher, and 
others, confirm the conclusion reached through chronogenesis. Moreover, the 
application by Beecher of the law of morphogenesis, as defined by Hyatt, 
and the recognition and establishment of certain primary characters have 
resulted in the discovery of a fundamental structure of general application for 
the classification of these organisms. It has for its foundation the nature of 
the pedicle opening and the stages of shell growth. On this basis Beecher 
(1891) has divided the class into four orders: the Atremata, Neotremata, Pro- 
tremata, and Telotremata. 
The nature of the pedicle opening being employed for ordinal divisions, 
persistent internal characters of the shell are, as a rule, used for superfamily 
purposes. Such are the presence or absence of a spondylium, brachial supports, 
etc. Family divisions are based upon a combination of external and internal 
generic characters, such as the outer form, nature, and position of muscles, 
internal plates, ete. 
No division, however, has any value unless the group contains forms of 
but one phylum, since a phylum or line of descent cannot originate twice. 
However, it happens that the same or nearly the same combination of mature 
characters is developed along different lines; and when this occurs the 
ontogeny will show it. It is therefore not correct to group different stocks 
under one and the same genus. For instance, the family TYerebratellidae 
probably divided during early Mesozoic times, one stock drifting into boreal 
and another into austral regions. These two stocks agree str ucturally i in the 
earliest shelled condition and also at maturity ; but between these two stages 
