Experiments were made to compare the practices of planting yam at the onset of the dry season (November) or at the beginning of the rains (February). Setts were also planted in February after pre-sprouting either outdoors from November/December, or indoors without water from November. Comparisons were also made between plots that were harvested first in mid-season and again at the end of the season (doubleharvesting), and plots that were harvested only at the end of the season (singleharvesting).

Setts that were pre-sprouted outdoors had low percentage emergence, produced weak plants, and yielded poorly whether double-harvested or single-harvested. Setts prepared and planted in the field in November emerged over a very long time, and attained half-maximum emergence only 10–19 days earlier than setts prepared and planted in February. Whether in double-harvesting or in single-harvesting, tuber yield/sett for the November planting was significantly greater than for pre-sprouted plantings, but not greater than for the direct February planting. The first harvest of the November planting had more mature tubers and a higher percentage of marketable tubers than the others. In all treatments, single-harvesting resulted in slightly greater yield than double-harvesting, but the differences were not significant. Each second-harvest tuber from double-harvesting was more amorphous in shape, more woody, and had more eyes (buds) on it than the single-harvest tuber. It is concluded that for efficiency of land utilization, February planting is preferable to November planting, and that singleharvesting is likely to be more economical than double-harvesting.

The effects of temperature and method of feeding on the intake characteristics and digestibilities of two milk substitutes were examined.

Forty Finnish Landrace x Dorset Horn lambs (20 males and 20 females) were fed one of two milk substitutes from 3 to 25 days of age when they were slaughtered and chemical composition of the empty body was measured.

The two milk substitutes contained 30% fat in the dry matter as butter fat (diet B), or tallow and coconut oil (diet TC).

The diets were offered at either 34°C (W) or 5 °C (C) and three methods of feeding were used. The diets were given: ad libitumcold (AL1), four feeds to appetite per day either warm or cold (AL2W and AL2C) and four feeds restricted to an intake of 62 g D.M./kg live weight0·5 either warm or cold (RW or RC).

Digestibilities of the milk substitutes were not affected by the temperature at which they were offered or the feeding regime used. Diet B had an apparent dry-matter digestibility of 97·6% and diet TC 92·5%. The fat digestibility of diet TC was 13·6 units lower than diet B and this affected the digestibility of all the other dietary components.

Lambs tended to consume more of diet B and grow faster.

Lambs on treatment AL1 consumed 21% more milk substitute than those on treatment AL2.

The composition of the body-weight gain was found to alter with rate of gain.

The temperature at which the milk substitutes were offered had only a minor effect on the performance of the lambs. The type of milk substitute and feeding regime used had a greater influence on both lamb intake and growth.

The experiment was on permanent grass at Rothamsted from 1974–75. ‘Nitro-Chalk’ (ammonium nitrate-calcium carbonate mixture, 25% N) broadcast in equal amounts for each of three cuts was the standard treatment against which the following were compared: (1) prilled urea (46% N) broadcast like ‘Nitro-Chalk’, (2) aqueous ammonia (26% N) injected in winter in bands 10 cm deep and either 30 or 60 cm apart, and (3) aqueous urea (18% N) injected like the aqueous ammonia. All these fertilizers were applied to give 250 or 375 or 500 kg N/ha in 1974 and all were reapplied in 1975.

In 1974 yields were larger with ‘Nitro-Chalk’ or prilled urea than with the aqueous fertilizers. Aqueous urea gave slightly larger yields than aqueous ammonia. Yields with the aqueous fertilizers were slightly larger when the injector knives were spaced 60 instead of 30 cm apart. Distribution of yield over three cuts was more uniform with ‘Nitro-Chalk’ and prilled urea (divided between cuts) than with the aqueous fertilizers. Injecting the aqueous fertilizers in bands 60 instead of 30 cm apart improved yield distribution only a little. In 1975 yields were slightly larger with the aqueous fertilizers at the first cut, but not in the drought afterwards.

Relative yields over the 2 years (‘Nitro-Chalk’ = 100%) were: prilled urea, 97; aqueous urea in 60 cm bands, 96; aqueous urea in 30 cm bands, 95; aqueous ammonia in 60 cm bands, 92; aqueous ammonia in 30 cm bands, 90.

In 1974 the grass recovered most N from ‘Nitro-Chalk’ and least from aqueous ammonia, but more from aqueous than from prilled urea. Recovery of N from the aqueous fertilizers was larger when injected in bands 60 instead of 30 cm apart, especially when 250 or 375 kg N/ha was given. Little rain fell in summer 1975 and recoveries of N were smaller than in 1974. The grass recovered more of the injected than the broadcast N, but no more with the injector knives 60 instead of 30 cm apart.

A total of 76 cows in seven, trials over a 3-year period were given a supplement of 53 g magnesiiun phosphate (supplying 11 g Mg and 10 g P/day) in 1 kg of concentrate for 21 days after starting to graze spring pasture. In contrast with 79 comparable cows given no supplement, there were significant improvements in the concentrations of magnesium and phosphorus in the blood of the cows, and significant reductions in the incidence of individual cows with low values.

Four young females of four ruminant species, namely cattle, buffalo, sheep and goat, were fed, in chopped form, a mixture of species of dry grass (Apluda aristata and Themada quadvivalvis; 3·85% crude protein) during a 28-day experimental period. Voluntary intake of grass was determined during the last 10 days and digestibility of feed nutrients during the last 6 days.

The digestibility coefficients of dry matter (D.M.) and crude fibre were higher for goats than other species, of organic matter, crude protein (CP) and ether extract were higher for goats and sheep than cows and buffaloes and of nitrogen-free extract higher for goats than cows and buffaloes and for sheep and buffaloes than cows. The daily intake of D.M., digestible CP and total digestible nutrients (TDN) is discussed in relation to body weight (BW) and kg W0·75.

Dry grass fed with mineral and vitamin supplement was adequate to sustain the body weights of buffaloes and goats but not of cows and sheep. Buffaloes utilized the dry grass better than other species as is evident from the highest feed and TDN intakes and marginal gain in weight.

Effects of phosphorus fertilizer, initial profile water storage and seasonal rainfall on yield and water use by dryland wheat on loamy-sand and sandy-loam soils were studied for a 2-year period (1973–5). Plant available water at seeding varied from 62 to 205 mm, seasonal rainfall varied from 62 to 154 mm and available P status ranged from 9·5 to 18·5 kg P/ha. Four rates of phosphorus 0, 15, 30 and 45 kg P2O5/ha were tested. Available water at seeding, seasonal precipitation and the available P status of soil determined the yield response to phosphorus fertilizer. Response to P application was observed up to 15 kg and 30 kg P2O5/ha in loamy-sand and sandy-loam soils respectively. Growth and yield of wheat were highly correlated with the available water at planting plus seasonal rainfall. Total water use did not change in loamy-sand soil, but it increased by 16 mm in sandy-loam soil with the application of phosphorus. The profile water depletion pattern, further, indicates that the fertilized crop used more water from layers below 135 cm in loamy-sand and 22·5 cm in sandy-loam soil.

A consideration of input-output relationships relating energy intake to its probable utilization for growth, condition and lactation by dairy cows leads to a series of equations describing observable responses. Energy is expressed in terms of units of milk equivalent, common to all equations, and the lactation curve, yn = anò e−cn is shown to be capable of biological interpretation, having regard to the subordinate equations from which it is derived. This expression relates milk production yn during a time interval n through the parameters a, b and c. The validity of the steps by which the function is developed is examined by reference to published work.

Observations were made of the ages at first appearance of each pair of permanent incisors (‘eruption age’), and the time from first appearance to full eruption of each pair (‘eruption period’), for Large East African Zebu cattle (Boran). For females the eruption ages were found to be 108·5±9·9 weeks, 137·2±12·9, 165·7±15·8, and 200·9±18·2 for incisor pairs one to four respectively; for steers the figures were 104·3±7·9, 130·3±10·6, 154·6±13·2, and 180·9±14·0; and for bulls 103·3±8·6, 127·6±11·2, 151·0±12·9, and 181·2±15·7. Females were significantly different from steers (P < 0.01), but steers not significantly different from bulls, for all four incisor pairs.

The distributions of eruption ages for each sex and pair can be considered as Normal for statistical purposes. There is a tendency for eruption to be asymmetrical after the first pair, with the right tooth preceding the left.

Correlations of eruption ages between adjacent pairs are of the order of 0·80 for all pairs and sexes; they decrease progressively as pairs intervene to 0·55 for the first and fourth pairs. Partial correlations approach zero when any intermediate pair is held constant. These results indicate that the whole eruption process is highly interdependent.

Evaluation of age grouping and estimations from the presence of incisors showed that this technique, although limited on its own, is still the best developmental trait for ageing. A steer with one incisor pair present had a probability of 0·108 of being the same age or older than one with two pairs. The ages of steers with one and two incisor pairs present were 115·7±12·05 and 140·7±14·27 weeks respectively.

Heritabilities of eruption age, for the various incisor pairs, varied between 0·91±0·33 and 0·43±0·25 for females, and 0·87±0·37 and 0·11±0·59 for steers.

Sources of environmental variation examined were variation between and within years, both of which were of the same order and a very small proportion of the total variation. The maternal environment had a small effect, probably specifically associated with milk consumption, which had a correlation of —0·22 for females and —0·12 for steers (P < 0.01) with the eruption age of the first incisor pair.

Eruption periods for females were 5·1±2·8 weeks, 5·7±3·8, 6·8±4·5, and 8·8±5·6 for pairs one to four respectively; for steers 4·6±2·6, 4·9±2·9, 5·6±3·8, 7·2±5·3; and for bulls 4·9±2·2, 4·8±2·7, 5·7±3·8, 6·0±3·3. Females were significantly slower than steers (P < 0·05) for the first three incisor pairs only, while steers did not differ significantly from bulls for any pair.

The distribution of eruption periods was positively skewed, and the correlations between pairs ranged between 0·0 and 0·35. Correlations of eruption ages and periods within incisor pairs lay between 0·50 and 0·95.

Variation of eruption periods was not significant within years, but was between years, and arose almost entirely from the first two experimental years. The majority of the heritabilities approached zero.

Growth rate differences probably accounted for the differences in eruption periods between the four incisor pairs and the three sexes. The cause of asymmetry of eruption was not the source of the variation in eruption times within incisor pairs, the first tooth consistently having a greater eruption time.

Ninety-six Friesian cows were used to determine the effects of pattern of allocation of a fixed amount of concentrates on milk yield during the treatment and residual periods (weeks 1–20 and 21–44 respectively) and on the whole lactation. Thirty-two were in their second and the remainder in their third or later lactation. Second-lactation cows were used to form separate blocks.

Four treatments were imposed for the first 20 weeks of lactation; these were Graded (G), Woodman (W), High Fixed (HF), and Low Fixed (LF). The first three treatments involved the same total amount of concentrate but fed in different daily amounts. Treatment W received Woodman's standard (Evans, 1960) amounts of feed according to a predetermined lactation curve derived from earlier experience in the herd; treatment G reallocated the same allowance by feeding more than W in weeks 1–12 and less during weeks 13–20; the HF treatment was a fixed daily amount of 6·4 kg, in total also equal to W total. LF was a fixed daily amount of 4·0 kg/day. The maintenance allowance was the same for all cows within a block and based on hay, dried sugar-beet pulp and barley meal. During weeks 21–44 all cows within a block received the same treatment; either grazing or self-fed silage and supplementary concentrates according to the time of year.

There were significant differences in milk yield during the treatment period resulting from plane of feeding (W, G and HF greater than LF; P < 0·001) but no significant differences arising from the pattern of feeding. In the residual period (weeks 21–44) there was no significant difference between the yields of second-lactation treatment groups. Thus for the whole lactation the yields of treatments G, W and HF were significantly greater than LF. However, for the older cows there were differences in the residual period and treatment groups G and LF gave significantly more milk than group W (P < 0·001 and P < 0·05 respectively). Over the whole lactation treatment group G gave significantly more milk than LF (P < 0·001), W (P < 0·01) and HF (P < 0·05). Furthermore, the increase in total lactation yield of G compared with W was 4·5 times the increase observed in weeks 1–20.

Cows of both age groups on treatment HF, unlike those on treatments W and G, never achieved a distinct peak yield but their mean weekly rate of decline was significantly less.

Live-weight changes varied between treatments and between age groups and were significantly correlated with milk yield in weeks 1–20.

There were no significant effects of treatments on milk composition in either age group at any stage of the lactation. The minimum value for solids-not-fat content for the cows on the LF treatment was unacceptably low.

It is suggested that for cows of moderate yield potential fed to Woodman's standards (Evans, 1960) the actual pattern of feeding a predetermined amount of concentrates has little effect on the yield or quality of milk. For cows of high potential a redistribution of concentrates to allow earlier more generous feeding will increase milk yield.

The rate of decline of milk yield is a characteristic of both cow potential and pattern of feeding and an adoption of a standard 2½% a week should be viewed with caution.

Seventeen gnotobiotic lambs were reared up to 21 weeks of age on cows’ milk followed by sterile solid diets similar to diets fed to conventional lambs. Seven were given limited defined populations of rumen bacteria, seven were left uninoculated and three were dosed with rumen contents from conventional sheep (‘conventionalized’). The lambs were reared in four groups corresponding to four lambing seasons.

Gnotobiotic lambs fed cows’ milk ad libitumgrew at similar rates and converted feed into live-weight gain with similar efficiency to conventional lambs. The gnotobiotic lambs irrespective of inoculation and the two conventionalized lambs in the first three groups failed to maintain growth on solid diets when milk feeding had been stopped. However, three gnotobiotic lambs in the fourth group, two of which were inoculated, continued to grow satisfactorily for at least 6 weeks on a solid diet alone, and the inoculation of defined populations of bacteria appeared to confer an advantage. The conventionalized lamb in the fourth group suffered a severe temporary setback, following which it grew well.

A comparison was made of the growth curves for early and late eruptors (steers whose eruption age of the first incisor pair was less than the mean, and greater than the mean respectively). From birth to 195 weeks early eruptors were consistently heavier than late eruptors (P < 0·01) by approximately 4% of the overall mean body weight. Growth was linear from birth up to 126 weeks of age, and the variability increased markedly from about 60 weeks of age.

The average daily gain from birth to 195 weeks was 341 g. Linear growth-rates for consecutive stages within this period were generally faster for early eruptors, with the differences ranging between 4 and 6% of the mean for periods exceeding 60 weeks. Comparison of these differences showed that linear growth-rates were only significantly different for growth periods of approximately 120 weeks and longer. When the specific growth-rates were compared no differences were significant.

The differences between the growth rates of fast and slow eruptors (steers whose eruption period was less than the mean, and greater than the mean respectively) were not significant, both for linear and specific growth rates.

Early eruptors had larger heads (head length x head width) relative to body weight (P < 0·05), indicating eruption at an earlier stage of body development. Also early eruptors had longer narrower heads, suggesting slower maturation rates.

Partial correlations between eruption age and head length varied from 0·37 to 0·74 (P < 0·01) showing a close relation with stage of head development. Partial correlations with other measures of skeletal development showed that they were of negligible importance relative to head development.

Relations between eruption age of the first incisor pair and six measures of carcass composition were negligible, as none of the partial regression coefficients was significant.

The effects of irrigation and nitrogen on S. 23 perennial ryegrass grown for seed were investigated in a series of field experiments from 1972 to 1974. Irrigation significantly increased seed yield by 16% in 1972 and 52% in 1974 but had no effect in the wet year of 1973. In 1972 maximum deficit reached 110 mm at the end of July and coincided with anthesis. Consequently the yield response was due to an increase in number of seeds per unit area and no other seed yield component was affected. In 1974 peak deficit also reached about 100 mm but started to build up rapidly very early in the season and had reached 80 mm by the time that the first ears emerged. Consequently the yield response was due to increases in number of fertile tillers, number of seeds per unit area and 1000-seed weight.

Irrigation had no significant effect on number of florets or seeds, except in 1974 when percentage of florets which produced seed was increased by 2%.

Irrigation had some effect on threshed straw yields, total dry matter, harvest index and total number of tillers but where this occurred the response was much smaller than that of seed yield which indicates that irrigation had greater effects on the reproductive development of the crop than on yield of dry matter and tillering patterns. Increasing the quantity of nitrogen from 0 to 80 kg/ha increased seed yields, all seed yield components except 1000-seed weight, threshed straw yields and total dry matter and number of tillers at most sampling dates. Increasing the quantity of nitrogen from 80 to 160 kg/ha had little further effect on the above components except in 1972 where seed yields were significantly decreased.

In 1972 number of florets was increased and percentage of florets which produced seed decreased with increasing quantities of nitrogen.

Comparisons were made between cultivated safflower varieties (Carthamus tinctorius L.) and six wild ecotypes (C. oxyacantha Bieb.) for their tolerance to simulated drought conditions. Plant characteristics of the wild strains were noted and soil samples from their collection sites were analysed for the electrical conductivity of the saturation extract. Polyethylene glycol 6000, mannitol and sodium chloride were used as osmotic substrates to prepare aqueous solutions having 0, 1, 2, 5, 8, 12, 15 and 18 atm. of osmotic potentials (OP). The influences of treatments were studied on seed germination, shoot length and fresh and dry weights of shoots and roots of seedlings.

Increased OPs progressively delayed and reduced seed germination, shoot length and fresh and dry weights of seedlings. Shoots were, however, more adversely affected than roots. Wild ecotypes were apparently more sensitive to high OPs than the cultivated varieties.

Birthooat fibre types were investigated in Barki, Merino and their crosses. The halo hair grade was high in the Barki breed, in the Merino a wide range of grades occurred but the majority was low. A trend of a decrease in halo-hair grade and an increase in non-hairy birthcoats occurred with the increase of Merino proportion in the crosses. An anterior reduction in halo-hair density was encountered in all groups under study.

Fibre type arrays were all coarse in Barki and ¼ Merino, the ¾ Merino had 50% fine arrays and 75% of the Merino arrays were fine. A trend towards a maternal influence on the birthcoat was observed in the ⅜ Merino and ⅝ Merino; more fine arrays were encountered in lambs born to dams with more Merino proportion.

Within saddle and ravine arrays, CT/Pre-CT ratio was highly significantly correlated with the pre-natal check as indicated by Merino proportion.

It was suggested, from the birthcoat studies, that ⅝ Merino might be the best combination in the Merino crosses as far as the fibre type arrays were concerned.

Less coarse arrays were encountered in the birthcoat due to selection against kemp in rams used for breeding.

The effects of irrigation and nitrogen on the soil-water regimes developed by S. 23 perennial ryegrass grown for seed were investigated in 1972 and 1974 by neutron scattering. Boots extracted water from a depth of about 105 cm by early May in 1972 and by early April in 1974. Extraction from greater depths was insignificant in both years. Without irrigation the maximum deficit was about 110 mm in both years. Deficits built up rapidly in 1974 but in 1972 were delayed to the end of the growing season. Irrigation limited maximum deficits in the top 35 cm of soil to less than 25 mm whereas without irrigation deficits reached as high as 40 mm in both years.

Transpiration from unirrigated grass was close to that calculated from meteorological data in 1972, but in 1974 measured crop transpiration was only about 60 %of the potential throughout the growing season. Roots of grass, whilst undergoing inflorescence initiation, apparently may be impaired in their efficiency of water uptake by deficits as little as 30 mm. For irrigated grass agreement between measured and calculated crop transpiration was poor, principally because of surface run-off. Nitrogen in the absence of irrigation appeared to enhance the rate of crop transpiration in 1972 but comparison with 1974 suggested that this effect was not real. Because of surface run-off resulting in uneven distribution of water, and the restrictions to crop transpiration imposed by limited soil water reserves, actual deficits should be monitored in irrigation experiments rather than relying on estimated values.

The rate of isotopic distribution of labelled phosphorus, added as H332PO4, between inorganic and organic phosphates and phosphates contained in micro-organisms was measured in pig slurry. Incorporation of 32P in all these phosphates occurred quickly in both aerated and non-aerated pig slurry. On the basis of a simplified model, turnover times for phosphorus were calculated to be of the order of 10–20 weeks for both non-aerated and aerated pig slurry.

Pig slurry contains 1–2% P (of dry matter) of which 10–30% is in organic molecules and 2–3% is in micro-organisms. About 10–20% of the organic phosphates is in solution, amounting to 10–20 mg P/l. The concentration of inorganic P in solution is of the order of 10–100 mg/1 though, at low Ca/P ratios in the feed, can be as high as 1000 mg/1.

Organic phosphates in solution in pig slurry are of high molecular weight and probably consist of DNA complexes with polyphosphates, Ca and (if used in the feed) Cu.

It is concluded that all organic phosphates in pig slurry are of microbial origin and that the feed composition has little influence on the organic phosphate content of the slurry.

Arguments for application of the results to pig slurry in general and to wastes from other animals are given.

The birth weight of Friesian and Holstein calves in Western Nigeria was generally lower than that recorded in other tropical countries. This was attributed to a low plane of nutrition during pregnancy and a consequent reduction in length of gestation of the dams. On average, pure-bred Friesian, pure-bred Holstein and Holstein x Friesian calves weighed 29·1, 30·8 and 30·7 kg respectively at birth. Male and female Friesian, Holstein and cross-bred calves averagely weighed 29·9 and 28·2 kg, 29·9 and 31·7 kg, and 32·6 and 28·8 kg respectively. Both the sex differences and breed x sex interaction were not significant.

Positive correlation coefficients of 0·45, 0·38 and 0·26 found between calf birth weight and gestation length for the Friesians (P < 0·05), Holsteins (P < 0·01) and their crosses (P > 0·05) respectively suggest that gestation length had a significant effect on Friesian and Holstein calf birth weight. In many cases, parturition occurred 8–16 days earlier than expected.

Calves born during the rainy season (April–September) were approximately 10% heavier at birth than those born during the dry season (October–March). The heaviest calves were born during the second half (July–September) of the rainy season. The breed x sex x season of calving interaction was significant for the Holstein breed (P < 0·05). Holstein male calves born during the rainy season were significantly (P <; 0·05) heavier than those born during the dry season. Friesian dams with three or more parities and Holstein dams that were 3 years old and above produced calves that were significantly (P < 0·01) heavier at birth than calves of younger dams.

A field experiment was made on Gurdaspur clay loam to assess the relative efficiency of various sources of fertilizer phosphorus for a paddy-wheat cropping sequence. Urea ammonium phosphate, superphosphate, a nitrophosphate fertilizer containing 30% water-soluble P and fertilizers containing 50 and 70% water-soluble P prepared from the nitrophosphate and superphosphate, dicalcium phosphate and rock phosphate were applied to the paddy crop at the rate of 0, 15, 30, 45 and 60 kg p2O6/ha and to wheat at double these rates.

The cumulative yield data for paddy and wheat show a slight reduction in efficiency of phosphatio fertilizers, with the exception of dicalcium phosphate, as their watersoluble content decreased below 70%. In general, the relative efficiency of the less water-soluble sources and of rock phosphate was considerably higher for paddy than for wheat.

The results show that the relative efficiency of low water-soluble phosphate fertilizers increased with time. The efficiency of phosphorus fertilizers declined with the rate of their application, more so in the case of higher water-soluble materials.

Barley plants normally produce some tillers which die at an early stage, without bearing an ear. To investigate the proposition that these non-ear-bearing tillers may be wasteful of plant resources, two experiments were made in which the effect of tiller growth on the growth and final size of the main shoot was assessed. In one experiment, tillers were removed at a very early stage when they were a few mm long, or later when they were just emerging from the subtending leaf sheath. In the other experiment the main shoot was removed from the embryo and the size of the coleoptile tiller which grew in its absence was measured.

Tiller removal affected the growth and final size of the main shoot. Leaves emerged more quickly on the main shoot of the detillered plants and were bigger and sometimes more in number. The main shoot of the detillered plants was heavier and it bore a greater weight of grain because the ear had more and heavier grains. In the experiments where the main shoot was removed the coleoptile tiller produced more leaves, which emerged more rapidly and it bore more grains than the coleoptile tiller of the intact plants.

It is concluded that tillers, during their initiation and early growth, compete with the main shoot for a limited supply of resources, thus reducing the size of the main shoot. As some of the resources of the plant are used to produce tillers which die at an early stage and make no contribution to the final grain yield, it appears that they should be regarded as wasteful and that their production may reduce the final grain yield. The physiology of the initiation, growth and early death of these non-ear-bearing tillers requires further investigation.

Five cultivation systems for spring-sown barley were tested on a clay to sandy-clay loam for 4 years and their cumulative effects on soil and crop measured. Four primary cultivations in the autumn, shallow (7·5 cm), medium (15 cm) and deep (23 cm) tine cultivation and mouldboard ploughing (23 cm deep) followed in spring by secondary cultivations and sowing, were compared with direct drilling seed into uncultivated soil after controlling weeds with herbicides. At the same rates of seed and fertilizer for all systems, direct drilling and shallow cultivation yielded significantly less grain than deeper tillage in the second year only; over 4 years the differences in mean yield were small and non-significant. In terms of crop output per unit of implement net energy (estimated) direct drilling was eight times more efficient than ploughing.

Direct drilling and shallow cultivation caused concentration gradients of ‘available’ phosphorus and potassium to form in the top soil. Soil strength and to a lesser extent bulk density were greater without than with cultivation but these changes had only a relatively small effect on root distribution in the profile. In untilled soil the proportion of the root system was less in the surface 2·5–12·5 cm and greater in the 12·5–32·5 cm horizon than in shallow or deep tilled soil.

Mean nitrogen uptake by the crop in the third and fourth seasons was least in direct - drilled barley probably due to lower rates of mineralization of soil nitrogen in compacted than tilled soil. It was deduced that for spring barley the ‘yield plateau’ was the same for each system but that optimal rates of nitrogen would be higher for crops in untilled than tilled soils.