The first problem beyond the stage of Haldane soup is imagining how protein polymers could form in dilute aqueous solution, when polymerization is a dehydration, or water-removing, process. Equilibrium strongly favors cleavage, not polymerization.

Sidney Fox has found that dry amino acids, heated to 160-210°C, will form polymers of molecular weights up to 300,000, provided that aspartic and glutamic acids are included in the mixture.

The sequences of these "thermal proteinoids" are not completely random, but show some internal order. These polymers display a limited catalytic activity, probably resulting from their charged side chains of acidic and basic amino acids. They catalyze the decomposition of glucose reasonably well.

It is important not to read too much into this catalytic activity, since even protons and platinum atoms are catalysts. It would be surprising if a polymer with such a mixture of side chains was not catalytic for some reaction. However, some such weakly catalytic polypeptides, with or without metal ions, probably were the ancestors of the much more efficient and selective enzyme catalysts of today.

These thermal proteinoids have another interesting property. If a hot proteinoid mixture is washed with water or salt solution, microspheres of a fairly uniform 20,000-Å diameter are formed, as in the photograph to the right.

These are small globules of proteinoid polymer solution, enclosed by a semipermeable proteinoid film with some of the physical properties of simple cell membranes.

Microspheres shrink and swell in salt solutions of different concentrations. They will grow at the expense of dissolved proteinoid material, and have been observed to bud like yeast cells to produce "daughter" microspheres. They can be induced to fission by MgCl2 or by a pH change. The enclosing film is a double layer resembling those found in soap films and artificial and natural membranes.