When I joined the Editorial Board of the Journal in 1949, I regarded the other five members as at least personal acquaintances and some as long-standing friends and colleagues. Indeed, apart from T. H. Middleton, I had known personally all the Editors from the founders in 1905, while my position on the staff of the School of Agriculture at Cambridge since 1931 and my close working association with R. H. Biffen and H. Hunter meant that the Journal occupied a special position in day-to-day life. Service on the Board since 1949 has provided the opportunity for what in effect seems a continuous association over a large part of the Journal's life commencing as it did with R. H. Biffen, A. D. Hall and T. B. Wood, who was the first Chief Editor, succeeded in turn by R. H. Biffen, F. H. A. Marshall, H. Hunter and J. Hammond, all of whom were my senior colleagues.

This paper describes statistical methods for analysing data arising from distinctness, uniformity and stability variety trials carried out to fulfil Statutory Regulations in the United Kingdom. Both univariate and multivariate methods are considered. It is shown that multivariate techniques can separate some variety pairs which could not be separated using univariate methods. A computer system for routinely analysing these trials is also described.

The utilization by sheep of dietary N provided in high protein, high water content fresh herbages (Ruanui perennial ryegrass, Tama Westerwolds ryegrass, Pitau white clover, and Fakir giant sainfoin at two stages of maturity) was studied at two levels of intake (maintenance and 1·5 maintenance). Feed was provided by constant feeder.

Apparent digestibility of N was similar for all herbages (ca. 85%) except sainfoin which, particularly at a late stage of maturity, was lower (70–80%). A small loss of nitrogen across the stomachs occurred with clover (1–3 g/day) and Tama ryegrass at the higher feeding level (2 g/day), but no loss was observed with the other diets. The apparent digestibility of N and of non-ammonia N (NAN) in the intestines was lower for sainfoin, and estimated true digestibility was also lower. Amino acid N contributed less to the NAN reaching the duodenum on the sainfoin diets than on the grass and clover diets.

N retention was negative at the lower level of feeding for ryegrass and clover diets. It was greatest for the sainfoin diets at similar N intakes, so that efficiency of retention of apparently digested N was also greatest for sainfoin.

The size of the urea pool, the plasma urea concentration and the urea irreversible loss, using [14C]urea, did not differ significantly between diets at similar N intake. Urea irreversible loss exceeded urinary urea excretion by 35—50% on all but the late-maturity sainfoin diet, where urea irreversible loss was more than double the urinary urea output. These data indicate dietary differences in the extent of degradation of urea on recycling to the gastro-intestinal tract. Urea clearance across the kidney was also lowest for sainfoin.

Data are compared in a simple model which illustrates the importance of variable clearance of urea across the kidney and the gut wall and the need for knowledge of factors which control this.

The first lactation records of 684 Sahiwal, 157 Brown Swiss × Sahiwal (BS × S) and 255 Friesian × Sahiwal (F × S) constituted the data for this study. The 300-day yield was estimated from part lactation yields of 13, 17, 21, 26 and 30 weeks post partum by four methods, namely, regression, ratio, gamma and inverse polynomial functions. The duration of the part record and the method of estimation were evaluated for their reliability in estimating the 300-day yield and sire evaluation.

The correlations between part records and 300-day yield ranged from 0·76 to 0·95, thecorrelations being higher in F × S. The error, absolute error and standard deviation of error weresmaller for ratio and regression methods than for the gamma and inverse polynomial functions. The part records of 30 weeks or more were found to be necessary to estimate 300-day yield within 15% of the actual yield with regression or ratio methods.

The rank correlations and product moment correlations between sire indices calculated with part records and 300-day yield were found to be highly significant. The part records of 17 weeks or more extended by ratio or regression could be used to evaluate sires as accurately as with 300-day yield by increasing the progeny group size by 7–10% in Sahiwal and by about 20% in F × S. Theloss of efficiency in sire evaluation amounted to about 10–15% with part records of 17 weeks or more compared with 300-day yield.

Indian Chokla and Malpura and their crosses with exotic Rambouillet and pure bred Rambouillet sheep were exposed to increasing temperature (20, 25, 30, 35 and 42 °C), or to 40 °C for 6 ha day for about a week, or suddenly to 40·4 or 44·2 °C to study their relative heat tolerance. Pronounced increase in the frequency of respiration under hot conditions together with increase in rectal temperature (under 44·2 °C) of Rambouillet contrasted with the physiological responses of the native breeds of sheep. Under thermal stress the Chokla and Malpura consumed more feed than the Rambouillet or the half breds. The Rambouillet × Chokla or the Rambouillet × Malpurahalf breds were apparently more heat tolerant than the Rambouillet sheep.

Seventy-five maiden Border Leicester × Merino ewes were divided into three treatment groups of 25: 1N, injected with PMSG on day 13 of a natural oestrous cycle in autumn (May); 1S, injected with PMSG on withdrawal of a progestagen sponge inserted for 13 days in autumn; 2S, treated as 1S but in spring (November). Five ewes in each treatment received each of 0, 200, 400, 800 and 1600 i.u. PMSG (dose, D). From 16 h after PMSG injection until oestrus, detected by entire rams, all ewes were bled from the jugular vein at 4 h intervals and plasma oestradiol-17β (E2) concentrations measured by radioimmunoassay. Time to onset of oestrus was recorded. Samples for estimation of plasma progesterone were taken on days 6 and 13 of the subsequent oestrous cycle. The ewes were laparotomized on day 5 or 6 of the cycle and all active corpora lutea, cystic follicles > 8 mm (most of which were luteinized) and follicles > 5 mm in diameter were recorded; the sum of these was taken as the total follicular stimulation at oestrus.

There were significant (P < 0·001) linear dose-responses to PMSG in terms of corpora lutea, total follicular development, progesterone, and E2. The covariance of E2 on total follicles was significant (P < 0·001). Progesterone concentrations were more highly correlated with the sum of corpora lutea and cystic follicles than with corporalutea alone (τ = 0·80 ν 0·73, both with 73 D.F.; P < 0·001). The time to onset of oestrus for treatment 1N ewes was highly variable due to the relative asynchrony of the time of PMSG injection but, when treatments 1S and 2S were analysed together, there were significant linear (P < 0·001) and quadratic (P < 0·01) effects of dose of PMSG, resulting in a decrease of 14 h in time to onset between the 0 and 1600 i.u. treatments. The covariance with E2 was significant (P < 0·05).

There was no difference between treatments in total follicular development but the number of corpora lutea was less (P < 0·05) in treatment 1N ewes than in the two sponge treatments (1S and 2S), again reflecting the asynchrony of treatment. There was no difference in progesterone concentration between treatments, but peak E2 was significantly less (P < 0·05) in treatment 2S (spring) than in 1N and 1S (autumn). Since there was no difference in the numbers of follicles which developed it appears that follicles in spring-treated ewes secreted oestrogen less intensively than in autumn. The lower oestrogen concentration was associated with a longer interval to onset of oestrus; linear regressions of time on dose PMSG (reciprocal) for treatments 1S and 2S separately had a common slope but significantly different intercepts (P < 0·01), so that oestrus was delayed by a mean of 6·4 h at all PMSG doses in spring compared with autumn.

Of the 73 ewes alive at lambing, 48 (66%) lambed producing 69 lambs. There was no significant effect of treatment on percentage of ewes which lambed but the lambing rate of those ewes with large numbers of ovulations (7–14) tended to be lower than that of the remainder (P ≃ 0·1). Of the nine ewes in this category, only four lambed and failure to lamb was significantly associated with excessively high plasma E2 (P < 0·05).

A method for the measurement of the net rate of synthesis of rumen microbes, entailing separate infusions of 35S-sulphate into the blood and rumen, is described and compared with estimates derived using digesta flow markers. Measurements were made on seven occasions with sheep and on eight occasions with cattle given chopped hay or pelleted hay diets at intervals of 60 or 90 min. Additional measurements were made on two sheep and two cattle given lucerne hay at intervals of 6 or 12 h. Estimates of microbial synthesis using the proposed method for animals fed at intervals of 60 or 90 min were 0·95–1·23 (mean 1·05) that derived using flow markers. For animals fed at intervals of 6 or 12 h the ratio was more variable, 0·91–1·35 (mean 1·17). The agreement between methods tended to be poor in animals with a low rate of irreversible loss of serum sulphate. The proposed method can be accomplished with a low analytical load, and appears to have potential for application in animals fitted with rumen fistulae only, and fed at infrequent intervals.

Blood samples were taken at 9, 10 and 11 weeks of age from 230 male Hereford × Friesian calves, the progeny of 12 Hereford bulls. Concentrations of blood glucose, serum albumin, inorganic phosphate and sodium were measured and correlated with body weights and growth rates until slaughter at 19½ months of age.

Correlations between growth rates and glucose concentrations (0·44) and between growth rates and albumin concentrations (0·38) were observed until 4 months of age. Similar correlations between body weights and blood glucose or albumin concentrations persisted until 6 months of age. By 10 months, however, the correlations had begun to decline, and by slaughter they had become insignificant.

Correlations between growth rates and inorganic phosphate or sodium concentrations were smaller, and also decreased with age.

The effect of hypoglycaemia on growth rate was compared with the effect of enzootic pneumonia. Although the 10% of calves with the lowest glucose concentrations were growing 24% more slowly than the other calves at the time of sampling, this growth depression was not related to pneumonia, and they subsequently made up for most of the early loss of growth.

Three blood samples were also taken from 22 Hereford bulls. No significant correlations were observed between the blood composition of the bulls and the body weights or weight gains of their progeny.

Body-weight gain, protein efficiency ratio (PER), standardized protein efficiency ratio (Std PER) and gross protein value (GPV, GPV1 and GPV2) bioassays were compared with nitrogen incorporation efficiency (NTE) for their accuracy and precision to determine protein quality of fish meal (FM), groundnut cake (GN), mustard cake (MS) and cottonseed cake (CS) for poultry. Results showed that all the bioassays were significantly (P < 0·05) correlated with NIE and were equally effective in ranking the samples of FM, GN, MS and CS according to their protein quality although body-weight gain procedure was less sensitive.

The present study was made to evaluate the protein quality of 11 samples of fish meal (FM) covering a wide range of nutrient content to find a suitable laboratory method for predicting nutritive quality. The nutrient content of the samples ranged from 34 to 62% in crude protein, 28 to 48% in true protein and 0·99 to 2·87% in available lysine. The true protein content of FM was further partitioned into myogen, myosin and stroma proteins. The stroma protein accounted for 61% of total crude protein suggesting that a large portion of FM comprised scales, viscera, cartilage, bones and other inedible parts of fish (head, tail, etc.). Neither true protein nor its different fractions was found suitable to detect differences in protein quality of FM although there was an inverse relationship between stroma protein and nitrogen solubility.

Available lysine content and relative nutritive value (RNV) were found accurate and reliable laboratory tests for assessing the protein quality of FM. The linear equations to predict gross protein value (GPV) from available lysine content and RNV are

GPV = 11·2×% available lysine+34·5;

GPV = 0·55×RNV+51.

The correlation coefficients between biologically assayed and predicted GPV from available lysine and RNV were 0·71 and 0·62, respectively. The findings suggest that the screening of FM samples for their nutrient content is essential before compounding diets for different categories of chickens.

A study was made to evaluate the protein quality of groundnut cake (GN), mustard cake (MS) and cottonseed cake (CS) and to find a suitable laboratory test for predicting the protein quality of these cakes for chicks. Different oilseed cakes varied considerably in their crude protein, true protein and crude fibre content. The nitrogen solubility was higher in GN (87·7%) followed by MS (78·7%) and CS (60·9%). A similar trend was evident with regard to in vitro protein digestibility. The GN protein was, however, lower in available lysine (2·54%) than MS (3·97%) or CS (3·19%). The albumin, globulin, prolamin and glutelin fractions accounted for 74·1, 57·0 and 48·0% in GN, MS and CS, respectively. The nitrogen solubility, globulin content and modified relative nutritive value (RNV) correlated (P < 0·05) with gross protein value (GPV). The accuracy and precision of various regression equations for predicting the quality of protein in GN, MS and CS are discussed.

The length and weight per unit length of individual internodes of mature reproductive tillers of S. 24 and S. 23 perennial ryegrass and S. 345 and S. 143 cocksfoot were examined under three manurial treatments. At the lower level of nitrogen application successive internodes generally increased in length from the base upward. With additional nitrogen there was an increased proportion of tillers in which internodes near the base of the stem were longer than those immediately above them; lateheading tillers were most affected.

Nitrogen increased the weight per unit length of upper internodes of all four varieties; the effect was most marked in S. 23 ryegrass. In the two ryegrasses, nitrogen reduced the weight per unit length of the basal internode.

At maturity the upper internodes had lower dry-matter digestibility values than the lower. Nitrogen reduced the dry-matter digestibility of upper and lower internodes.

The results are discussed in relation to selection criteria in variety synthesis.

Cowpea is an important food legume crop of arid and semi-arid regions of the tropics. In such climates whether a determinate or indeterminate type of growth habit would be more useful is not clear. In the present study a determinate and indeterminate cultivar of cowpea were grown at two population densities under rainfed conditions for 2 years. Grain yield was higher in the indeterminate variety for both years. At higher density more dry matter was produced but it had no effect on grain yield. Nitrogen analysis showed that it moved from leaves to developing pods. However, a large amount of nitrogen was still left in vegetative parts in contrast to cereals. Pod development was completed in 19 days and the rate of dry-matter accumulation during the peak period of growth was as high as 120 mg/day per fruit. It was difficult for the plant to cope with this high demand for photosynthates. It is suggested that more pods can develop on a plant provided the growth rate of individual pods is slower and extended to a longer period.

The effect of ammonium ions on artificially imposed water flux and trans-root potential difference was studied in excised sunflower roots. Water movement through the root system decreased and the potential of the xylem sap became less negative with respect to the external medium as the ammonium ion concentration in the external medium increased. It is suggested that ammonium ions appeared to inhibit water uptake either wholly or partially through a general or specific blockage of root metabolism which reflected on the permeability of water through root membranes. The reduction in the trans-root potential was due to the effect of ammonium ions on the original trans-root potential of epidermal root cells. This may indicate that the site of water uptake inhibition lies within the root epidermis.

Large populations of bees often continue to forage on oil-seed rape crops until flowering ceases. Applications of insecticide to crops in the late flowering stage could be hazardous to these bees.

We present results from an experiment at Saxmundham measuring the effects of growing different numbers of successive winter wheat crops, each year, from 1971 to 1976; it followed five cropping sequences introduced from 1966 to 1970.

From 1971 to 1973 four seed rates, two row spacings and three nitrogen rates were tested in all combinations with four of the previous cropping sequences. The remaining sequence was sown traditionally for continuity. Number of ears, leaf area, yield and nutrient content showed little benefit from increasing or decreasing seed rate, around that usually sown. Rows 15 cm apart gave the largest yield independently of N, because they produced most ears. There was no loss in yield from growing up to eight consecutive crops of winter wheat. 100 kg N/ha was sufficient for maximum yield.

From 1974 to 1976 two varieties of different potential (Cappelle-Desprez and Maris Huntsman), chlormequat chloride (CCC), three amounts of N, as single, or as divided dressings, were tested in all combinations with the four cropping sequences. The fifth sequence again was sown traditionally, but with or without autumn-applied N. Maris Huntsman gave a grain yield only 0·22 t/ha larger (4·5%) than that of Cappelle-Desprez, and mean yields were larger (also by 4·5%) where CCC was given. Divided dressings of N increased yields by 0·11 t/ha (2·2%). The mean increase in yield from adopting all three improvements was 0·5 t/ha (or 11%). Yields of the 10th and 11th consecutive wheats declined, probably because black grass (Alopecurus myosuroides) became abundant. There was a good inverse relationship between losses of N in drainage during winter and subsequent yield and grain nitrogen uptake, which helped to explain large yield variations from year to year.

The rate of leaf extension (LER) in swards of perennial and Italian ryegrass was examined in relation to daily mean air temperature. Cultivars studied were an Italian, an Italian-perennial hybrid, a tetraploid perennial and diploid perennial cultivars in three maturity groups. A range of response to temperature was apparent. The Italian, the hybrid and the tetraploid perennial cultivar showed higher LER values than those of the diploid perennials. Amongst perennial cultivars the tetraploid cv. Reveille was shown to have a lower rate of photorespiration than the diploid cv. Gremie. These data are discussed in relation to the use of daily temperature observations in predicting leaf growth.

Domesticated cowpeas (Vigna unguiculata (L.) Walp.) have larger seeds than their wild relatives. The effects of this increase in seed size on seedling emergence and growth were investigated. In shallow sowings, to 5 cm, seed size had no effect on emergence, but in deeper plantings ahigher proportion of the larger-seeded accessions emerged, and they emerged earlier than smaller-seeded accessions. The size of seedlings just after emergence was inversely related to depth of planting and directly related to seed size. The greater size of the seedlings of larger-seeded accessions was maintained until at least 40 days after sowing, despite the higher relative growth rates of smaller, wild seedlings. When compared at the same size there was no general difference in the growth rate of wild accessions and that of domesticates. During the evolution of domesticated cowpeas, large seeds have probably been advantageous because of their better emergence and establishment under agricultural conditions, and because of the greater size of plants arising from them.

A model for predicting the various transformations undergone in the soil by nitrogen applied in farm wastes, in response to variations in soil and climatic factors, is presented. The soil is divided into layers and a simple mass balance model is used to describe the movement and redistribution of water within the soil profile, as a function of rainfall, evapotranspiration and soil moisture characteristics.

Ammonification of native and applied organic N is assumed to follow first-order reaction kinetics and nitrification and denitrification are treated as zero-order reactions. The rates or rate constants for all three reactions were related separately to soil temperature, moisture content and pH.

Uptake of nitrogen by a grass crop, dry-matter production, removal of dry matter and N through harvests, addition to soil of C and N through dead plant material and stubble and the consequent immobilization of mineral N are allowed for. Some of the predictions of the model are compared with the results of a lysimeter experiment.

An assessment was made of the relative merits in the short term of KH2PO4and pig slurry as sources of orthophosphate. This involved a comparison of these two sources of P with regard to (1) adsorption of P by the soil from solution, (2) maintenance of the solid-phase P in labile form and (3) rate of release of the solid-phase P into solution.

The adsorption isotherms made with two soils revealed that the amount of orthophosphate adsorbed from pig slurry was much higher than that from KH2PO4 at all levels of application. Conversely, in order to raise the orthophosphate concentration in the soil solution to a given level, much larger applications of orthophosphate would be required as pig slurry than that in the form of KH2PO4. Therefore, the risk of direct orthophosphate-P pollution of ground water from pig slurry is likely to be even less than that from soluble P fertilizers.

Determinations of exchangeable-P in treated soils after different times from treatment showed that a larger fraction of the solid-phase P derived from slurry orthophosphate remains in an exchangeable form than that from KH2PO4.

The rate of desorption of solid-phase P into solution measured using an anion exchange resin was higher in the KH2PO4-treated soil than that in the slurry-treated soil at different times from addition of equal amounts of orthophosphate-P.