236 HETEROCHRONY AND PAEDOMORPHOSIS

grasses; when effective concentrations are reached, mitosis proceeds equally over the entire body. Since the liter of juvenile hormone decreases steadily throughout larval ontogeny, the effective concentration is reached later and later in each successive molt, thus permitting a longer allometric phase in each instar and the eventual production of an adult (ontogeny following the last larval molt is entirely allometric, corresponding to the first phase of other instars). Novak (1966, p. 116) emphasizes that the normal relation of growth and development can be dissociated, to produce, for example, giant supernumerary larvae geometrically similar to smaller, normal forms.

W. M. Krogman, the leading American student of child growth, makes the same distinctions between growth, development, and maturation. He defines growth as "proportionate changes in size," and adds:

At the risk of oversimplification it is possible to set up three aspects of human growth and development.

1. We grow. This is size . . .

2. We grow up. This is proportion.

3. We grow older. This is maturation. Every tissue in the body bears an indelible register of the passage of biological time. (1972, p. 3)

Dissociation of the Three Processes

Attempts to separate simple growth from changes in shape date at least from Aristotle's speculation that the embryo receives two kinds of food from its mother, one to generate form, the other to increase size: "Everywhere the nutriment may be divided into two kinds, the first and the second; the former is 'nutritious,' being that which gives its essence both to the world and to the parts; the latter is concerned with growth, being that which causes quantitative increase" (Historia animalium, 744b, 32–36). Needham (1933) presents a fascinating array of separations between growth and development, ranging from "anidian" embryos that never develop a primitive streak and continue to cleave without differentiating, to the experimental production of dwarfs geometrically similar to their larger forebears. Polyploid larvae of salamanders have fewer and larger cells than normal forms, but they build the same internal structures with the same size and shape (Goss, 1964, pp. 31–33). Whatever controls the development of form is at least partly independent of the size and number of building blocks. Wolpert points out that this ability of a system to maintain its form as parts are added or removed (in fact, to exhibit general size in-