"After the rapid glance," he says, "at these variations, which indicate to us the special changes undergone by each species, there arises a more important consideration, and the view of which is broader; it is that of the transformation (_changement_) of the species themselves; it is that more ancient modification which has gone on from time immemorial, which seems to have been made in each family or, if we prefer, in each of the genera in which were comprised more or less allied species."[141] In the beginning of his first volume he states "that we can descend by almost imperceptible degrees from the most perfect creature to the most formless matter--from the most highly organized animal to the most entirely inorganic substance. We will recognize this gradation as the great work of nature; and we will observe it not only as regards size and form, but also in respect of movements and in the successive generations of every species." "Hence," he continues, "arises the difficulty of arriving at any perfect system or method in dealing either with nature as a whole or even with any single one of her subdivisions. The gradations are so subtle that we are often obliged to make arbitrary divisions. Nature knows nothing about our classifications, and does not choose to lend herself to them without reasons. We therefore see a number of intermediate species and objects which it is very hard to classify, and which of necessity derange our system, whatever it may be."[142] This is all true, and was probably felt by Buffon's predecessors, but it does not imply that he thought these forms had descended from one another. "In thus comparing," he adds, "all the animals, and placing them each in its proper genus, we shall find that the two hundred species whose history we have given may be reduced to a quite small number of families or principal sources from which it is not impossible that all the others may have issued."[143] He then establishes, on the one hand, nine species which he regarded as isolated, and, on the other, fifteen principal genera, primitive sources or, as we would say, ancestral forms, from which he derived all the animals (mammals) known to him. Hence he believed that he could derive the dog, the jackal, the wolf, and the fox from a single one of these four species; yet he remarks, _per contra_, in 1753: