moiety measuring 31.85 μ. Each of these portions was again subdivided by a septum; the distal one, by a septum B, into a distal moiety just over 17 μ in length, and a proximal moiety not quite 13 μ long; the proximal one, by a septum B', into a distal moiety measuring just over 17 μ, and a proximal moiety slightly more than 15.5 μ. Close observation showed that each of these was again sub-divided by a faint, approximately median septum, into two cells.

Now it is perfectly obvious from this that the various subsidiary segments or cells, measured at any period, are not of exactly the same length.

At 12.27 I again measured, as exactly as possible, the relative positions on the scale of the above septa to see if I could determine the relative rates of growth of the various segments in the intervening twenty minutes.

The septum A had moved backwards nearly 3 μ on the scale, and the apex and base of the whole segment were driven apart about this distance. This suggests that the growth (intercalary) had occurred chiefly in the distal part of the original segment, and such was the

case.

The distance between the septum B and the apex was not measurably altered, i.e., the part to the distal side still measured 17.29 μ; but the distance between B and A had increased from 12.74 to 16:38 p. On the proximal side of the septum A no appreciable changes had occurred. It is true, my measurements, as multiplied out, are a trifle different from those first made, but the differences may be neglected, as of course I could not measure to the decimals concerned.

What does come out, and very clearly, is that during the interval named the growth was entirely in one only of the four cells, and not in all of them. It is obvious that such phenomena complicate the question of the growth of the entire filament still more, and no doubt some of the minute variations observed are due to such events as these.

I made the following measurements under the 1/20th immersion, on a terminal segment 36 μ long, in the sister culture to the last one,

standing side by side with it. The temperature was rather low, 15° C., and the observations soon came to an end owing to the closefocussing objective cracking the very thin cover slip. I made many attempts to measure growths for longer periods under this lens, but the failures were so frequent owing to the extremely thin cover slips and hanging drops necessary for so high a power, and the difficulties of illumination, that I had to abandon them.

The value of each division of the micrometer scale was found to be from 1.20 μ to 1.25 μη but here again I found it hopeless to attain to greater exactitude of measurement. However, the following attempt is at least interesting.

On comparing the figures with other measurements at the same temperatures, moreover, they agree very well; so that it may yet be possible to carry out measurements with this power.

It should be noticed—as a point of importance in what followsthat these minute variations observed with the highest powers are not traceable (and probably neutralise one another along the filaments) when longer stretches are measured with the lower powers.

After considerably greater experience with these curves, I am able to sum up the meaning of these experiments more clearly.

1. They give evidence that the growth of the filament as a whole is intercalary, and due to increase in length and division of all its cells, along the entire course of the filament.

2. The different rates in the general growth observed are due partly to differences in temperature, partly to differences in age of the portions observed, partly to differences in the food-medium in which the organism is growing, and partly to other causes.

3. The small variations in rate of growth, especially those traced under high powers, are due partly to small and unrecorded variations in temperature, e.g., cooling of the thin cover-slip when the bell-jar was lifted (as in the experiment on p. 306), and partly to the causes assigned on p. 302, namely, pauses during the intercalation of the new segment walls, and, no doubt, to some extent, to curvatures in the filaments, and want of practice on my part in recording the observations so accurately as I learned to do later on.

Comparative Measurements.

The foregoing results led to the attempts-now to be describedto grow two filaments side by side, one in the light and the other in the dark, to see if the action of light could be detected by any change in the growth curve.

Before passing to these experiments, I obtained satisfactory evidence that two cultures, side by side and under the same conditions, behave similarly.**

On February 18th I started the following comparative cultures, to see how far I could test the action of daylight rendered so probable by some of the foregoing results.

Spores were sown in broth at 9 A.M. in two hanging drop cultures, each with a layer of water at the bottom to prevent rapid changes of temperature or drying up. The two cultures were then put under a dark bell-jar, covered with foil and brown paper, at 18° C., rising to 21° C., close to the south window where I intended to start the experiment. The cultures were left thus so that the spores should have time to germinate out normally, which they began to do about noon; they were left undisturbed till 2 P.M.

Meanwhile, I had selected two similar bell-jars and placed these a the window, each over a microscope provided with a thermometer; one bell-jar was darkened with tin foil and brown paper, the other not. The thermometers were examined, and readings taken from time to time. The differences observed at first, and due to the shading action of the covering, began to get less and less as the bell-jars, table, and microscopes, were warmed by the sun, which was bright and hot; and by noon the temperatures were very nearly the same in each jar.

At 12.30 the thermometer under the light jar indicated a temperature of 25.5° C.; that under the darkened one =246° C.,† and this difference of nearly 1° C. was maintained till about 1.30, the readings being as follows:

About 2 P.M. the sun began to sink behind trees and the roof of a building facing the window, and now it was possible to use the bright light from the sky without danger of direct insolation.

*Better proofs of this were obtained later, however, and are given at pp. 359-361. † These are all air temperatures, unless the contrary is specially stated.

At 2.20 P.M. the thermometer in both bell-jars stood at 23° C., and, as the curve of temperature shows, they were so nearly equal throughout the experiment that I had little hesitation in concluding that the cultures did not vary much from the same temperature at any time; because I thought it could hardly be imagined that the hanging drops in these damp cells rapidly vary with the temperature of the environment, and still less so the filaments growing in them.

At 2.30 the observations were begun by noting the temperature, and measuring the rod in each case. As the following tabular record shows, the difference in time between the two notes rarely exceeded a couple of minutes, whence the two series are closely comparable in all respects.

It was not until much later that the question arose-or, rather, acquired the great importance I now attach to it-how far the culture in the light could avail itself of or be affected by the infra-red rays reflected from the mirrors, and so complicate the matter of temperature effects in these experiments. This matter is of pregnant importance in its bearing on all physiological experiments of this kind, however.