For case of comparison the data are presented in tabular form rather than as charts. (Tables 3, 4, 5). In these tables some of the columns require a word of explanation. The designation of each rat is made according to the following method: The first letter indicates the year in which the litter was used, B = 1927 in this case; the second letter indicates a certain litter born and used that year; the Roman numerals I, II, etc. indicate the groups into which that litter was subdivided. When necessary the individual rat in a group received

Fig. 1.—To illustrate the method of determining the relative values of effects on growth: XYZ = the “area of growth.”

a number, 1, 2, etc. The second column gives the rise in weight, i.e. the difference between the weight when the special dieting began and the maximum attained during the feeding or soon after it ceased. For greater accuracy the actual weights are given in brackets. The “area of growth” in the third column was obtained by measurement of the growth-curves as follows: As stated already, after the ten

days of oyster feeding the rats were kept on basal diet till they died. Before death the weights fell usually to a lower point than that at which the feeding began. A horizontal line was drawn across the curve at this level and the included area was measured carefully in sq. cm. by a planimeter such as surveyors use for evaluation of areas on maps.^* Figure 1 gives an example of this procedure. Rats Bu IV were given a meal containing oyster-spawn for ten days beginning at the point x; the maximum weight was at y, and at z the growth curve is back at the same level as x. The “area of growth” is the roughly triangular area xyz and in these cases measured 11.6 and 13.2 sq. cm. on the original chart. The advantage of this figure is that it indicates to some extent the amount of storage of vitamin-A that the rat was able to accomplish.

It is not claimed that this method can give a true mathematical measure of the amount of vitamin present in foodstuffs, but it gives a figure that can be used for comparative purposes.

The next column “Prolongation of life” gives in days the duration of life after the oyster feeding began. In the control groups (total = 30 rats) this figure averaged 20 days, and it will be noticed that the oyster feeding caused a distinct prolongation of life beyond this period whether growth had occurred or not. This figure may be regarded as an indication of a “maintenance” factor just as the “growth area” shows “growth + maintenance” factors.

The columns that refer to eye-symptoms need no further explanation, but note may be made of the fact that almost every case showed eye-trouble. That has been our experience in this laboratory during the past three years.

Comments on the Tables: Table 3 gives the results of individual feeding on undried oyster; this series was carried out late in the year, when all the samples had been received, and this made possible a certain amount of overlapping, i.e., the groups of one litter were given oyster material collected in different months. For these reasons the results are probably more reliable for comparative purposes than the others, and this method would have been adopted to a greater extent if the samples had all been available simultaneously, and if the best level of doseage had been known, for a great many experiments were done on the March, April, and May oysters at various levels ranging from 0.5 grm. to 3.0 grms. in order to determine the best level, viz., a dose that lay between the minimal and the just maximal. In the litter Bs three pairs (buck and doe) were fed March, May, and July oyster respectively. The area of growth and the prolongation of life indicate progressive falling off in vitamin-A content as the season advanced.

In litter Bt two pairs similarly received June and August oysters, and again there is evidence of less vitamin-A in August as compared to the earlier month, but it will be noticed Bt I showed a better result with June oyster than Bs III did on May oyster. This is probably due to a difference in the quality of the litter and not to a greater content of vitamin-A in the June as compared to the May oyster.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

The only other striking feature is the marked improvement in the September and October oysters as compared to the July and August ones. At the 1 grm. level these are nearly as rich as any of the others at the 2 grm. level—this is shown by the “area of growth,” by the prolongation of life, and the effect on the eye-symptoms, and similar results can be seen in Tables 4 and 5 when dried material was used.

In Tables 4 and 5 the only groups that “overlap” are Bu II on July oysters and Bu I on August oysters, where again the former proved the better sample, and Bt III on unspawned October oyster and Bt IV spawned oyster where weight for weight there was surprisingly little difference. There is clearly a tendency to low results in June, July, and August with a fairly high figure for March and May and a very high figure for September.

Tinned Oysters: A sample of tinned Stewart Island oysters was also examined. In making the leaven, care was taken to make the oyster solids correspond to the average of the fresh oyster. As in the case of the tinned toheroa the vitamin content was very high; thus the two rats Bu III (Table 3) fed on this material showed a much better result than their litter-mates, Bu II, fed on the same amount of May oysters. The canning of these oysters was said to have been carried out late in the season of 1926 but the exact date could not be ascertained.

Discussion: While these experiments show that when fed at the 2 grm. level these Stewart Island oysters nearly always show a high content of vitamin-A, the writer feels that the results of comparison of the oysters from month to month leaves much to be desired. If the comparison is again undertaken he would suggest that more of the “overlapping” method should be followed and that the method of comminution adopted by Jones, Murphy, and Nelson (4) should be followed—(grinding while frozen.) as probably better than either the “leaven” or the “meal.” As stated at the beginning of this paper it was hoped that some relationship between the gross chemical composition and the vitamin-A content might be found, but apparently such is not the case. The September oysters, though rich in vitamin, show no special peculiarity in composition, except for the fact that they were the heaviest of all. There are two likely sources of variation in the content of vitamin-A that occur to one—firstly, the oysters' food—secondly, the presence of the spawn: in regard to the first, September is the Spring in this, the Southern, Hemisphere, and with an increase in the intensity of sunlight there is likely to be a more or less rapid increase in the plankton on which the oysters feed, but in the same connection one has to remember that the successive samples could not be dredged always from exactly the same ground. The trawlers usually work the ground from east to west and there is said to be a considerable difference in the nature of the bottom and therefore probably of the food between the two extremes of the shoal.

Secondly, as to the growth of the spawn—this occurs steadily all through the winter up to the time of spawning (end of October). At this time there is a drop in the percentage of glycogen, but the vitamin content of oysters ripe for spawning and washed free of the