The analysis of ordinary bacteria shews them to consist[69] of about 85% of water, and 15% of solids; while the solid residue of vegetable protoplasm contains about one part in a thousand of sulphur. We may assume, therefore, that the living protoplasm contains about 1/1000 × 15/100 = 15 × 10^{−5} parts of sulphur, taking the total weight as = 1. But our little micrococcus, of 0·15 µ in diameter, would, if it were spherical, have a volume of π/6 × 0·15^3 µ = 18 × 10^{−4} cubic microns; {41} and therefore (taking its density as equal to that of water), a weight of 18 × 10^{−4} × 10^{−9} = 18 × 10^{−13} mgm. But of this total weight, the sulphur represents only 18 × 10^{−13} × 15 × 10^{−5} = 27 × 10^{−17} mgm. And if we divide this by the weight of an atom of sulphur, we have (27 × 10^{−17}) ÷ (275 × 10^{−22}) = 10,000, or thereby. According to this estimate, then, our little _Micrococcus progrediens_ should contain only about 10,000 atoms of sulphur, an element indispensable to its protoplasmic constitution; and it follows that an organism of one-tenth the diameter of our micrococcus would only contain 10 sulphur-atoms, and therefore only ten chemical “molecules” or units of protoplasm! It may be open to doubt whether the presence of sulphur be really essential to the constitution of the proteid or “protoplasmic” molecule; but Errera gives us yet another illustration of a similar kind, which is free from this objection or dubiety. The molecule of albumin, as is generally agreed, can scarcely be less than a thousand times the size of that of such an element as sulphur: according to one particular determination[70], serum albumin has a constitution corresponding to a molecular weight of 10,166, and even this may be far short of the true complexity of a typical albuminoid molecule. The weight of such a molecule is 8·6 × 10166 × 10^{−22} = 8·7 × 10^{−18} mgm. Now the bacteria contain about 14% of albuminoids, these constituting by far the greater part of the dry residue; and therefore (from equation (5)), the weight of albumin in our micrococcus is about 14/100 × 18 × 10^{−13} = 2·5 × 10^{−13} mgm. If we divide this weight by that which we have arrived at as the weight of an albumin molecule, we have (2·5 × 10^{−13}) ÷ (8·7 × 10^{−18}) = 2·9 × 10^{−4}, in other words, our micrococcus apparently contains something less than 30,000 molecules of albumin. {42}