Let's keep going with the embryogenesis metaphor. It makes it really easy to see the difference between the two types of repetition that Deleuze has now uncovered:
it is essential to break down the notion of causality in order to distinguish two types of repetition: one which concerns only the overall, abstract effect, and the other which concerns the acting cause. One is a static repetition, the other is dynamic. One results from the work, but the other is like the 'evolution' of a bodily movement. One refers back to a single concept, which leaves only an external difference between the ordinary instances of a figure; the other is the repetition of an internal difference which it incorporates in each of its moments, and carries from one distinctive point to another.
Our bodies appear to be made of repeated forms like our hands, but the identity of these are merely the final steps in a dynamic process of differentiation that is repeated in each of our limbs. Before, when we thought about repetition we were continually asking how a particular form could be identically repeated in a world of natural and moral law, or in a world where our conceptual representations should be able to capture any real difference. We took the finished identity of the form in question for granted, as if we had a little picture in our heads of one hand, and went looking for somewhere in time or space where we found another matching hand that duplicated it. Now, however, we're interested in the hidden repeated process that may (or may not) give rise to outwardly repeated forms.
What's changed, you might ask? Are we just shuffling terms around here, exchanging "process" for "form"? It's important to stop and wonder about this. The most important change is that the shift in level from form to process means we are no longer taking the form as pre-built and repetition as a game of matching, but are trying to explain how the form got built twice. We're actually explaining the repetition, rather than just noticing it (which only pushed the question of what's repeating off into another realm anyhow, because how did we know it was repeated? How did we become conscious of the repetition? Or how did the repetition become conscious of itself as such?).
Shifting levels has a bunch of other consequences though. Now that the repetition has a cause we are no longer engaged in a matching memory game, so the cause and the effect (the thing repeated) don't look anything like another another. Specifically, the genetic program that (often, in the right mileu) produces one hand and another hand doesn't look anything like a hand; there's no representation of a hand in the genome. There's no "gene for the hand" or picture of the hand in the embryo. We often hear about the "genetic blueprint". A worse metaphor could hardly be devised. The fascinating part of embryogenesis, and morphogenesis more generally, is precisely that there is not a picture of the finished form contained in the genome, not even "in code". The code just codes for proteins. The proteins interact with the code to produce new proteins that attach to one another and create chemical gradients that effect the concentrations of other proteins, etc ... There's no picture of the finished product nor of the intermediate states, there's just a cascade of differences.
Nor is there a pre-defined space to contain a pre-defined form. The embryo doesn't know "where" to put the hand. It's a bit like the old joke about how long a man's legs should be (long enough to reach the ground). Pre-existent forms require a simple location in a pre-existent space. With embryogenesis though, the space and the form grow together. In some sense you might even say that the two become indistinguishable, or at least go round and round in circles. This is an active space that isn't just a passive receptacle or container for finished forms.
in the dynamic order there is no representative concept, nor any figure represented in a pre-existing space. There is an Idea, and a pure dynamism which creates a corresponding space.
Ideas are eggs. Seeds from which stuff grows when the conditions are right. But also, if they are to be successful organisms, the result of interactions with those conditions.