The development is recorded of the series of experiments with potatoes at Rothamsted during 1925–27, designed to examine the quantitative response of yield to varying quantities of nitrogenous and potassic manures, and to test the relative value with this crop of different sources of potash.

While rather precise comparisons were obtained on the qualitative question by means of Latin squares in 1925–26, the reality of the depression ascribable to chloride could not be demonstrated in these years, but became clearly apparent when in the following year, the qualitative experiment was merged with the quantitative one.

In the earlier quantitative experiments, although satisfactory responses were obtained, the precision of the results left much to be desired, since only four replicates could be used. When by merging the experiments this was increased to nine replicates, much smaller responses were clearly measurable.

The large and complex type of experiment finally adopted thus supplied more precise information on both heads than could previously be obtained, and in addition to a more thorough exploration of the different combinations possible.

1. Cereal plots were sampled by three different methods; two systematic, and one involving a random location of sampling units.

2. The disadvantages of the systematic methods as compared with random sampling, emerged clearly.

3. These disadvantages were further emphasised in an analysis of earlier data on sampling methods. For this purpose the methods and. results of certain recent contributions to statistical theory were used.

4. By the use of a random sampling method, the variance due to sampling errors may be made a satisfactorily small fraction of the total variance of cereal plots one-fortieth of an acre in area.

The object of this series of investigations is to secure detailed information concerning the composition, digestibility and nutritive value of pasture grass in its different stages of growth. The results which were obtained in these respects by cutting the herbage of the experimental pasture plot at weekly and at fortnightly intervals have been described in previous communications. During the season of the present experiment, the trials have been carried a stage further by the adoption of a system of cutting at 3-weekly intervals. The main findings of the 1928 investigation are recorded below:

(1) Chemical composition of 3-weekly pasture cuts: The adoption of a more lenient system of cutting at 3-weekly intervals led to a slight lowering of the percentage of crude protein in the grass and a slight raising of the percentages of crude fibre and N-free extractives. On the other hand, no corresponding effect was noted in respect of the ether extract, SiO2-free ash, lime and phosphate, the percentages of these constituents being very similar in the weekly and 3-weekly pasture samples obtained in 1928. The falling off of the percentage of crude protein in the 1928 3-weekly-mown herbage, as compared with the weekly and fortnightly-mown herbage of 1925 and 1927 respectively, was not wholly the consequence of the more lenient system of cutting, but was also due in part to the protein-depressing influence of the droughty periods which were experienced in the 1928 season.

1. (a) Pyrethrin I and II have been isolated by the method of Staudinger and Ruzicka from the insecticidal plant Pyrethrum (Chrysanthemum cinerariaefolium). Both are shown to be highly toxic to the insect Aphis rumicis.

(b) Pyrethrin I was found to be the most toxic substance so far tested by us and, as it was about ten times as toxic to these insects as pyrethrin II, it is concluded that it is mainly responsible for the contact insecticidal value of pyrethrum.

2. Two micro-analytical methods of determining the pyrethrin content are described, (a) By means of the acids after hydrolysis, (b) By means of the semicarbazone. They are given on pp. 278, 282.

3. The analytical results obtained for a series of pyrethrum samples agreed with their observed insecticidal properties to Aphis rumicis.

4. Comparisons of the pyrethrin contents, as estimated, with the results of direct toxicity experiments both on the pyrethrum samples and the pure pyrethrins, confirm the validity of the analytical methods.

5. There was a significant and positive correlation, in the samples tested, between the amounts of pyrethrin I and II.

6. Insufficient data are available to show a significant correlation between the size of flower-heads and the content of poison, or to draw conclusions as to the effect of external conditions such as soil, weather or age of bed.

The investigation bears out the conclusion of the Rothamsted experiments that leaching out may occur any time throughout the winter up to March, or rarely April.

The loss of nitrates is directly limited by the amount of drainage and so is indirectly determined by the rainfall, and by temperature—as it affects the rate of evaporation.

All three nitrogenous fertilisers are shown to lose well over 50 per cent, of a normal dressing by leaching out during the winter, but the loss is greatest in the case of sulphate of ammonia, least in the case of rape dust, calcium cyanamide being intermediate.

The writers wish to express their indebtedness to Miss N. Sugg who carried out the experimental work in the first winter.

Experiments were carried out at two centres in Leicestershire during the spring of 1928, with the object of comparing the values of mangels and dried sugar beet pulp in the winter ration of dairy cows.

Two groups of cows were selected at each centre, and the usual method of feeding for a period on one ration and then reversing, was adopted. Composite samples of the milk from each group of cows were taken at six consecutive milkings per week, during the period of the experiment. Determinations of fat and solids not fat were made, and the yield of milk at each milking recorded.

Although there are indications that the change of ration caused a slight temporary variation in the quality of the milk, the secretion of milk solids followed, in a general way, the variation in yield: This is confirmed by the average composition figures representing the whole of the milk from each particular ration.

Thanks are due to Mr Thos. Hacking for placing every facility for obtaining these data at the disposal of the author.

Inconsistency in the returns of sugar percentages from the factories has often perplexed growers and has at times given rise to some mistrust of the analytical procedure. Consignments of beet lifted on the same day and from the same field have been credited by the factory with widely different sugar contents. This has occurred within the writer's experience, even though the precaution was taken of loading the cart loads of beet into two trucks alternately, thereby eliminating any possibility of one truck being filled with beet from a better part of the field than the other.

(1) It is proposed to estimate the amount of organic carbon in soils by determining the amount of sulphur dioxide produced in the ordinary Kjeldahl digestion. The gaseous products of reaction are passed through standard iodine solution, and the excess iodine titrated with standard sodium thiosulphate. Details of the method are given.

(2) The results obtained with a number of soils of differing character and origin are compared with the figures obtained for organic carbon by dry combustion. The sulphur dioxide method gives results which average 89.6 ±1.03 per cent, of the combustion figures. It is proposed therefore that the percentage of organic carbon found by this method should be corrected by the factor 100/89.6 = 1.116.

(3) The percentage recovery of carbon indicated by the proposed method is rather higher for pure substances but still falls short of 100 per cent.

(4) The proposed method is applicable to carbonate soils without the necessity for any correction for inorganic carbon.

(5) It is likely that soils containing inorganic reducing substances such, as sulphides will give high results by the proposed method.

(6) Absorbing the sulphur dioxide in 25 per cent, sodium bichromate, it is possible to determine the carbon dioxide by passing the gases through standard baryta in a Reiset tower. The organic carbon thus indicated agrees with that by the sulphur dioxide method.

(7) From data with certain peats, it appears that the factor 1.724 for converting organic carbon to organic matter is too low.

In the first paper of this series (10)1, Keen and the present writer discussed work by earlier authors on the approximate specification of the nature of a soil by a single soil constant, in place of a detailed analysis, and described experimental work designed to show the significance of a number of simple physical measurements. The chief objects of the present paper are to describe the results obtained with such measurements as applied to a number of Natal soils, and to discuss the value of some other easily obtained physical data as a means of specifying the nature of a soil. Work of this type was discussed at the International Congress of Soil Science at Washington in 1927, and it was resolved that co-operative work on an international basis should be undertaken.

1. The variations in production due to age have been investigated with the records of 738 Ayrshire cows for 4380 lactations.

2. Milk and butterfat production increase up to about 7 years of age and then show a decrease.

3. The fat percentage for 3-year-olds is higher than that for older cows.

4. After 3 years of age there is little change in the fat percentage with age that is of any practical significance until advanced ages are reached, when there may be a fall of importance.

5. The increase in production associated with age is probably attributable, in part, to the growth of the secretory tissue of the udder and to body growth in general.

6. Part of the increase may also be due to an improvement in functional activity through use.

7. The tendency for milk to show a slightly lower fat percentage as the cow advances in age is probably due to the fact that as the milk yield changes the fat yield changes in the same direction but at a slower rate.

8. There is little known regarding the influence of very advanced age on production, but it is probable that many cows maintain for a long time the production associated with maturity, and then decline slowly.

9. Heifers with a low fat percentage need not as a rule be expected to test higher on reaching maturity.

10. It is probable that the increase in production with maturity is associated more closely with high initial production than with persistency of production.

11. Correction factors for age are presented.

In a field trial with lucerne grown from seed treated with varying doses of culture it was found that the numbers of nodules were increased as the dose was raised from 2,500 to 20,000 organisms per seed (56 to 7 Ib. of seed per culture). Storing the seed for periods up to 28 days between inoculation and sowing, caused some loss in the nodule numbers. This loss was greatest between 1 and 7 days' storage.

The difference in dose of culture and in period of storage did not significantly affect the crop subsequently obtained from the inoculated plots, whose yield was, however, much above the uninoculated.

In a pot experiment made with runner beans, it was found that increase in the dose of culture above 1,280,000,000 organisms per pot containing six seeds was still capable of increasing nodule numbers but not to an extent proportional to the increase in dose.

The experiment does not exclude the possibility that the restriction in effect of very heavy doses may be due to the soil population becoming saturated with the bacteria. On the other hand, observations on lucerne plants grown aseptically on agar and inoculated with a pure culture, showed that even when excessive numbers of the bacteria immediately surrounded the root hairs, only 4 per cent, of these were infected.

(1) A large series of data on the duration of gestation in the goat has been analysed and the constants calculated.

(2) The frequency curve is unimodal and symmetrical, suggesting a single factor or a variety of factors acting simultaneously as the cause for birth.

(3) Slight breed differences exist in the duration of gestation in the goat as in other species examined.

(4) There is a continuous variation in the duration of gestation with the time of year at which conception occurs. Spring conceptions give shorter gestations than autumn conceptions.

(5) There is a distinct difference in the duration of gestation for young and older dams. This is related to the age of the dams and not to the order of the births. Gestation is shorter in the young animal than in the older.

(6) In the goat the size of litter has little or no effect on the duration of gestation.

(7) Constants for duration of gestation in other species have been obtained. Variability increases with the duration of oestrus and about half the variability in any species with a long oestrous period may be ascribed to this cause.

(8) The factors involved in fixing the duration of gestation in the species examined are evidently the same in all cases, and probably culminate in a single agency responsible for terminating gestation.

Very considerable amounts of money are annually expended in Mauritius for the purchase of artificial fertilisers, and consequently it is of importance to know what happens to these substances when they are added to the soil. Extensive laboratory experiments have been carried out with nitrogenous fertilisers(1, 2), and also with phosphatic manures(3), and in order to complete the series an investigation into the availability of potash when applied in various forms was undertaken. Local practice favours the use of potash chiefly in the form of nitrate of potash and molasses, and on this account the availability of the potassium oxide in these two substances was tested, and in addition, in potassium sulphate.