The term "heterochrony" refers to changes, over evolutionary time, in the rate or timing of developmental events. It is probably the most widely used concept relating development to evolution. The speeding up or slowing down of growth does appear to be a common occurrence in evolution, and can lead to very different looking adult animals. A precise definition of heterochrony has been hard to nail down, though, because any change in morphology must necessarily involve a change in the rate or timing of something in development. Thus, everything winds up looking like heterochrony - and the term looses its usefulness.
One way around this problem is to limit the term heterochrony to those cases in which a developmental process is uniformly sped up, slowed down, or has its onset or offset shifted in time. We thus exclude cases in which different parts of a developmental process change rates to different degrees, since this points to changes in the nature of developmental interactions, not just the rate at which something is happening.
Consider the ontogenetic trajectories shown below.
Those on the left represent a case in which one trajectory (the dashed line) is simply an accelerated version of the other. Both of these curves are produced by equations of the form:
with different values of the parameter (r). This is apparent when we transform the size axis so as to linearize the solid trajectory; the dashed trajectory is also linearized. In contrast, the trajectories on the right show a different kind of relationship. The equation producing the solid line is the same as above, but the dashed line is produced by:
In other words, there is a change in the structure of the dynamical system, not just a change in the rate at which it proceeds. Note that this also means that the two trajectories cannot be simultaneously linearized.
Calling changes such as that on the right heterochrony makes the term so general as to be useless. It also makes it impossible to assert that all heterochrony falls into one or another basic type, such as neoteny, hypermorphosis, and so on.
On the other hand, restricting the term heterochrony to cases such as those on the left above makes the concept both testable and biologically meaningful, since it corresponds to a specific class of transformations.
Defining heterochrony in this way we can test for it, given two ontogenetic trajectories, by asking whether we can superimpose one trajectory on the other by a uniform stretching or translation along the time axis. The transformations corresponding to heterochrony are illustrated below.
The first three cases correspond to the traditional "kinds" of heterochrony.
(In addition, two other terms are often used: Paramorphosis, in which the adult of the ancestor resembles a juvenile of the descendant, and Paedomorphosis, in which the adult of the descendant resembles a juvenile of the ancestor. These last two terms describe the outcome of development rather than the process itself.)
If such a transformation superimposes one trajectory on the other, then we can infer that one of the developmental processes could give rise to the other through a uniform rate change. If not, then there must have been some change in the structure of the developmental process, not just a change in the rate at which it proceeds.
As an example to show the value of such comparisons, consider the ontogenetic trajectories for brain size in humans and chimpanzees. The plots below show brain weight plotted against age, starting at conception (much of the interesting growth takes place before birth).
The plot on the left shows the untransformed trajectories, that on the right shows the result of transforming the chimpanzee data with sequential hypermorphosis. The plots below the trajectories show the distances between the chimp data points and the mean of the human data for the same age. Note that the transformed chimp trajectory matches the human trajectory quite closely and residuals of the transformed chimp data are evenly distributed around the mean of the human data. This implies that there have been no substantial changes in the process of brain growth since our common ancestor - just a uniform extension of each growth phase.
By contrast, compare the trajectories for humans and macaques, shown below.
Transforming the macaque data with sequential hypermorphosis does not bring it into line with the human data. A statistical test based on the length of "runs" of residuals on one side of the line leads us to reject this as an example of sequential hypermorphosis (similar analysis rejects all other forms of heterochrony in this case). We thus conclude that there has been some change in the nature of the growth process that occurred between our most recent common ancestor with old world monkeys (~25 mya) and our ancestor with chimps (~7 mya). This change appears to have been the addition of a novel growth phase, beginning at around the point of birth and continuing for about 1 year in humans (about 9 months in chimps).
A similar analysis reveals that squirrel monkeys (a new world monkey) have the same kind of trajectory as macaques. Combining all of these results, we can plot the changes in the form of the brain growth curve on a phylogeny of the primates for which we have data. This is shown below.
This tree shows that the major changes in the pattern of brain growth had taken place before our common ancestor with chimps; subsequent change on the human line being primarily an increase in overall size achieved by extending each growth phase.
We can also compare humans and chimps with respect to whole body growth. In this case, no one transformation works for the entire trajectory. Interestingly, though, if we break the body growth curve into two parts - before and after the onset of puberty - then each part is related to growth in chimps by a different combination of heterochronic transformations.
Transforming the chimp growth curve with a combination of sequential hypermorphosis and neoteny produces a very good match to the human growth curve up to the point of the onset of puberty; after which point the trajectories diverge. Interestingly, the curves after this point can be almost exactly matched by applying the same amount of sequential hypermorphosis as before, but with no neoteny.
Thus, relative to chimps, human growth shows an extention of each growth phase (sequential hypermorphosis). Furthermore, up until the beginning of adolescence, human growth is uniformly slowed down relative to chimps. Starting in adolescence, though, human growth accelerates to the same rate seen in chimps.
The combination of these factors results in humans having a much more distinct adolescent growth spurt that do our closest relatives. The figure shown above illustrates that, relative to our cousins, we have not adapted to grow rapidly during adolescence, but rather to grow slowly before it; thus streatching out our childhood.
For further discussion of the mathematics of ontogenetic trajectories, see:
Rice, S. H. 1997. The analysis of ontogenetic trajectories:
when a change in size or shape is not heterochrony. PNAS 94:907-
912.
Get a PDF reprint (232 Kb).
For further discussion of heterochrony in primate evolution, see:
Rice, S. H. 2002. The role of heterochrony in primate brain
evolution. In: Human evolution through developmental change. N. Minugh-Purvis
& K.J. McNamara, eds. Johns Hopkins Univ. Press, Baltimore.
Get a PDF reprint (650 Kb).
Return to Sean Rice's page