We now have some of the key concepts needed to describe Darwin's model. The non facit saltum axiom implies that Darwinian systems are resultant in the sense that any trajectory can be decomposed into smaller steps that have similar aggregate effects. However, Darwin believed that these continuous trajectories could produce surprising outcomes. Using language anachronistically we can say that Darwin was comfortable with time-asymmetry and innovation, but not with non-linearity. The systemic surprise had to emerge gradually.
Unlike Lamarck, whose evolutionary theory was broadly philosophical, Darwin devoted a lot of space in Origin to evidence-based description. He was aware that the fossil evidence did not line up. Georges Cuvier (1825), for example, had assembled a mass of empirical evidence that geological epochs were punctuated by rapid collapse and the emergence of a new type of ecosystem with a new fauna and flora (discussed by Huxley 1875-1889). Darwin played the taphonomic gambit familiar to any archaeologist. The fossil record was incomplete and patchy; the absence of evidence for missing links and linearity was not evidence of absence.
Darwin's commitment to non facit saltum may have been motivated by a fear of political theory. Darwin was a Whig, a scientist and a gentleman. Like all of his class, he would have been aware of, and appalled by, the French Reign of Terror. He would also have been uncomfortably aware of the Tory view that the Enlightenment and revolution were all of a piece, and that educating the masses was a recipe for insurrection. By insisting that the new system emerge from the old gradually, he was ensuring that his book could not be attacked as a pretext for revolution. By suggesting that natural selection could generate novelty, he acknowledged the possibility of gradual reform.