The use of weight gain in long-term experiments can produce serious bias in treatment comparisons unless the initial weights of treatment groups are equal or nearly so. When observations are repeated on the same experimental unit, covariance analysis of the data for each occasion produces treatment comparisons adjusted by the regression on the covariate at that time; polynomial trends in these adjusted comparisons with time can be easily calculated even though the regression on the covariate varies with time.

There exists a large body of soil-water theory for ‘modelling’ soil-water movement and distribution. Yet simpler and sometimes cruder methods are often preferred. In part this preference is due to the fact that theory is not yet able to solve completely the more complicated situations that occur naturally, such as the intermittent application or removal of water in the presence of very hysteric soil-water properties. But in many instances it is due to the apparent difficulty of the analysis for the non-mathematically minded. For example, descriptions have been given of soil-water models, based on direct physical arguments, that are in fact finite difference analyses of the governing partial differential equations, but which make no reference to such equations or their existence. Indeed, the impression is sometimes given that soil-water physicists produce unnecessary, irrelevant and complicated analyses for something that is really quite simple. Unfortunately, the simpler approaches bypass possible complications that are revealed by the mathematically based analyses. It is therefore the purpose of this paper to take advantage of the close relation between the two approaches, and to draw attention to the steps that may be taken to avoid or reduce possible difficulties when following the simpler methods. In particular, it is hoped that the paper will lead to a more efficient use of computing facilities.

Two experiments were conducted to study the influence of heat treatment of Vicia faba beans on the productive performance of laying hens fed diets containing large proportions of the beans. In the first experiment, the effects of extruding and pelleting the beans were studied while in the second experiment the effects of autoclaving at 121 °C for 30 min the cotyledons and hulls of the beans were evaluated. Hyline-W36 layers were used in the first experiment while two strains (Shaver-288 and Dekalb) of birds were used in the second experiment. Each experiment lasted 5 months during which data on egg production, egg weight, feed intake, mortality and body-weight gain of birds were collected. The results obtained indicated that heat treatment (extruding, pelleting and autoclaving) of the beans, cotyledons and hulls had no beneficial effect on the productive performance of laying hens fed on diets containing large proportions of the heated beans, cotyledons and hulls. Egg-size reducing factors appeared to be concentrated in the cotyledons rather than the hulls of the beans.

Three experiments each utilizing four buffalo heifers in 4 × 4 latin square designs were conducted over a period of 84 days to estimate the voluntary consumption of eight long-chopped and four ground roughages. Voluntary intake and rate of passage were determined during the last 10 days and digestibility of feed nutrients during the last 6 days of a 21-day experimental period.

The mean daily dry-matter intake (DMI, g/kg W0·75), D.M. digestibility (%) and mean retention time (h) of long and chopped foods were 81·6, 65·7, 75·2; 81·4, 54·2, 60·3; 78·0, 58·6, 66·0; 75·8, 58·5, 66·3; 67·5, 68·1, 87·9; 62·1, 40·9, 73·3; 56·3, 50·6, 70·9 and 55·5, 58·8, 82·4, respectively for maize silage, sorghum stover, dry grass, maize stover, berseem hay, wheat straw, pearl-millet stover and cowpea hay. Appreciable differences were observed between feeds in the digestibility of D.M., organic matter (OM) and crude protein (CP). No consistent pattern in level of intake and either CP or crude fibre (CF) contents of feeds or their D.M. digestibilities was discernible. The relationship between DMI and D.M. digestibility was low (r = 0·29). Higher DMI of feeds tended to be associated with faster rate of passage excepting maize silage and berseem hay. The relationship between D.M. digestibility and mean retention time was low (r = 0·46) and between total digestible nutrient intake and body weight was high (r = 0·79).

The mean daily DMI (g/kg W0·75), D.M. digestibility (%) and mean retention time (h) of ground roughages were 68·7, 51·3, 58·0; 62·9, 38·4, 60·0; 62·4, 40·1, 63·0; 53·1, 66·6, 72·0, respectively for ground maize stover, wheat straw, dry grass and berseem hay. Ground roughages appeared to be utilized less efficiently by buffaloes than longchopped roughages.

Cabbage accessions were screened for resistance to turnip mosaic (TuMV) and cauliflower mosaic (CaMV) viruses in glasshouse and field tests. None was immune to infection by either virus, but some were highly resistant to one or both viruses. TuMV induced severe necrotic lesions on the outer leaves of susceptible plants at harvest, and this was associated with severe internal necrosis in heads stored for 8 months at 0–1 °C. This correlation was not observed in CaMV infected plants. Pepper-spot necrosis was low in stored heads of some accessions but high in others.

Six experiments were carried out during 1973 and 1974 to study the effects of some cultural practices on tobacco. In Expts 1 and 2 effects of N, topping and suckering were studied. Leaf area was increased by N and topping in both experiments. Suckering significantly decreased, and slightly increased, the area in Expts 1 and 2, respectively. The number of leaves per plant was increased by topping and suckering in Expt 1, but decreased by both in Expt 2. N decreased the number of leaves in both experiments. Yield of cured leaf was increased by all three factors in Expt 1 and by suckering only in Expt 2, whereas both N and topping slightly decreased the yield in the latter experiment. Quality was significantly lowered by the higher rate of N, while generally improved by topping and suckering.

In Expts 3 and 4, effects of plant density, N and stage of suckering were studied. The yield of cured leaf was increased by the higher plant density, by 9 and 15%, and decreased by the higher dose of N, by 15 and 7%, in Expts 3 and 4, respectively. Early suckering was beneficial in Expt 3 (16%), but harmful in Expt 4 (8%). Leaf quality was decreased by the higher rate of N, but improved by the higher plant density. Time of removal of suckers had no significant effect on yield or quality.

In Expts 5 and 6 effects of supplementary irrigation at different stages of growth were studied. Both the number of leaves per plant and yield followed very similar trends. The best yield and quality were obtained when irrigation was not limiting during all three phases of growth. However, the last phase (C) was the most critical phase regarding irrigation in Expt 5, while the first two phases were the most critical ones in Expt 6, since withholding irrigation during phase C reduced yield by 44–50%, and only 6%, in the two experiments, respectively. Leaf quality seemed to be best when irrigation was applied throughout the growth period, particularly the first two phases of growth.

In 1974 three cultivars of barley of different straw lengths, Mazurka, Julia and Midas, were grown alone and with each of four cultivars of red and two of white clover in a split-plot randomized-block experiment of four replications. Although the barley cultivars differed significantly in yield the barley x clover interaction was not significant and yield of grain was not affected by the presence of the clovers.

In the following year the entire experimental area was sown with Mazurka barley. Significantly higher grain yields were obtained on those plots sown with clover in 1974 than in plots where no clover had been sown. In no case was the combined yield of barley from the 2 years on any treatment involving clover lower than that of the mean of the no-clover control plots, although the highest increase in yield, 11% from the S.123 red clover treatment, over the no-clover control was not shown to be significant.

Seven non-pregnant Murrah buffalo heifers aged 20–34 months and weighing 170–260 kg were used in a 7 × 4 incomplete latin square design to study the effect of feeding mixed diets on voluntary intake over a period of 84 days. Voluntary intake and rate of passage were determined during the last 10 days and digestibility of feed nutrients during the last 6 days of each 21-day experimental period.

Maize stover, dry grass and wheat straw were fed ad libitum mixed with either a fixed quantity of berseem hay (MH, GH and WH rations, respectively) or concentrates (MC, GC and WC rations, respectively) to total co. 0·80% of body weight (BW). Berseem hay fed free choice mixed with limited (0·70% BW) concentrates (HC ration) comprised another treatment. The mean daily total dry-matter intake (DMI g/kgW0·75) and total digestible nutrient intake (given in parenthesis) by animals were 62·2 (1·94), 72·2 (2·20), 74·3 (2·30), 75·2 (2·94), 79·5 (2·59), 85·9 (2·78) and 88·5 (2·98), respectively on MH, GH, WH, HC, MC, WC and GC rations. The digestible crude protein intake on the HC ration was higher than on the remaining six rations. Feeding a fixed quantity of either berseem hay or concentrates resulted in lower consumption of basal roughages suggesting that the physical capacity of the gut limited intake.

Dry-matter digestibility (%) and mean retention time (h), given in parenthesis, were 57·6 (69·4), 56·3 (64·6), 56·7 (67·5), 72·2 (79·9), 59·2 (620), 59·4 (620) and 58·9 (66·0), respectively on MH, GH, WH, HC, MC, GC and WC rations. The digestibilities of D.M. and proximate principles, except crude fibre (CF), of the HC ration were higher than those of the remaining six mixed diets. Concentrate feeding tended to hasten the movement of digesta of the roughage portion of the mixed diet through the digestive tract. The longest mean retention and rumen retention times were associated with the highest digestibilities. The wheat straw-berseem hay mixed diet was as good as dry grass or maize stover-berseem hay mixed diets with regard to its nutrient intake and weight gain by the heifers.

Data for 383 steer carcasses from the first 3 years of the Meat and Livestock Commission's Beef Breed Evaluation programme were used to evaluate a standardized commercial cutting technique for estimating breed differences in carcass composition. The sample included carcasses from cattle out of Friesian, Hereford × Friesian and Blue-Grey dams by sires of the main traditional British beef breeds and by Charolais, Limousin and Simmental sires. The left side of each carcass was divided into 14 standardized commercial joints which were deboned and trimmed of fat to a fixed level (commercial cutting); the other side was jointed in the same way and each joint fully separated into lean, subcutaneous fat, intermuscular fat and bone (full separation). Breed means for characteristics obtained by commercial cutting were adjusted to equal carcass subcutaneous fat percentage (SF) estimated by visual appraisal, while those for full separation characteristics were adjusted to equal SF obtained by tissue separation. Breed differences in deboned fat-trimmed joints (saleable meat) as a percentage of carcass weight were in good agreement with those for carcass lean percentage (between-breed correlation, rb = 0·92), as were those for saleable meat to bone ratio and lean to bone ratio (rb = 0·91). There was a poorer relationship between fat trim and separable fat (rb = 0·62) but there was little breed variation in these characteristics at equal SF. Breed differences in the percentage of total saleable meat distributed in higher-priced joints were very similar to those for lean distribution in higher-priced joints (rb = 0·87).

Sixty-one experiments with 15 levels of K fertilizer in the presence of excess N and P fertilizer were carried out on adjacent sites of the same field. Yield was always related to level of K fertilizer by a ‘diminishing returns’ type curve, and a derived equation, which defined relative responsiveness in terms of a single parameter, fitted the data for each crop very satisfactorily. Although the responsiveness of many of the crops was similar there were marked differences and the optimum levels of K (defined as the level at which a further 10 kg/ha increased yield by 1%) varied from 0 to 360 kg/ha, depending on the crop. Responsiveness was largely independent of the plant family to which the crop belonged, but was related to the mean plant weight atharvest; the larger the weight the less reponsive the crop. No general relation existed between responsiveness and duration of growth.

The % K in the dry matter of leaves (including stems) at harvest of crops receiving the optimum levels of K fertilizer was mainly determined by the family. It was generally between 0·9 and 1·1 for the Amaryllidaceae, between 1·1 and 1·2 for the Leguminosae and between 1·9 and 2·5% for the Cruciferae. The difference between the % K in the dry matter with the optimum level of K fertilizer and that with no fertilizer was proportional to responsiveness. Percentage K at harvest was a good indicator of the extent to which crop growth was restricted by lack of potassium.

At harvest crops receiving the optimum levels of K fertilizer contained between 29 and 220 kg/ha of K, but uptake increased asymptotically to a maximum as K applications were raised to higher levels. Maximum uptake for nearly all crops was almost double the uptake with the optimum fertilizer application.

Percentage recovery of 100 kg/ha of added K fertilizer varied between 8 and 70%, roughly in proportion to the total crop dry weight, which varied between 1 and 15 t/ha.

Effects of level of K fertilizer on crop quality were also measured and over the practical range of applications the effects were generally small.

The differences between the K requirements of crops are discussed and it is argued that the responsiveness of one crop relative to that of another would be expected to be similar on a range of soils.

Three experiments, each with six lactating Mehsana-Surti buffaloes, were conducted to investigate the effect of supplementary feeding of concentrates on intake of basal rations of berseem hay, maize silage or wheat straw and on the total intake of feeds in 3 × 3 latin square designs.

In the berseem hay experiment, concentrate feeding at 16·8 or 26·1% of hay drymatter intake (DMI) resulted in a decrease of hay DMI by 0·77 and 0·65 kg/kg concentrate D.M., respectively. Intake of total digestible nutrients (TDN) and digestible crude protein (DCP) of buffaloes receiving concentrates in addition to hay were not different from those receiving hay alone. The possibility of some chemical factor limiting feed intake has been indicated for these rations in buffaloes. Supplementary feeding had a little effect on milk production.

In the maize silage experiment, concentrate feeding at 14·5 and 27·7% of the silage DMI resulted in a similar increase of total DMI, thereby increasing the TDN intake and DCP intake. Appreciable increases in milk yield and solids-not-fat (SNF) content due to supplementation were noted, although milk-fat content tended to be low.

Voluntary intake of rations comprising wheat straw fed free choice with three levels of concentrates at 28·7, 48·0 and 68·8 % of straw DMI was studied in the third experiment. Concentrate supplementation had little effect on straw DMI with the result that intakes of total D.M., TDN and DCP on the medium and high proportions of supplements were higher than those on the low proportion. The milk yield, of buffaloes receiving medium and high proportions of concentrates with wheat straw increased significantly over those receiving the low proportion of concentrates. The SNF and milk-fat contents were similar on all the three treatments. The implication of physical factors limiting intake is discussed in the case of maize silage and wheat straw diets.

The lactating buffaloes failed to maintain their weights on either sole berseem hay or sole maize silage rations. Supplementary feeding helped buffaloes to register small weight gains on hay rations and substantial gains on silage rations. The buffaloes on wheat straw fed with three proportions of concentrates tended to put on a little weight. Multiple regression using pooled data of Expts 1 and 3 showed that the partial regression coefficient of metabolic body weight (kgW0·76) on TDN intake was not significant but that of 6% fat-corrected milk on TDN intake was, suggesting a close relationship between milk yield and food intake.

The interactions of breed, season and environmental temperature influencing reproductive activity in Indian sheep breeds have been studied so far in some detail only with respect to the physical characteristics of the semen (Sahni & Roy, 1969; Mittal & Ghosh, 1979). The literature on the influence of these factors in determining the concentration of some of the important chemical constituents of ram semen is scanty, except for two recent reports by Nivsarkar, Dwarkanath & Pant (1977, 1978) in which data on fructose and citric acid concentration in electro-ejaculated semen of Magra rams, obtained during the winter and spring, have been presented. A need was felt for understanding the pattern of concentration of these constituents throughout the year in semen collected from rams under natural conditions. The present investigation was, therefore, undertaken to study the seasonal pattern of concentration of two important chemical constituents (fructose and citric acid) and physical (ejaculate volume and sperm concentration) characteristics of the naturally ejaculated semen of two important Indian desert sheep breeds, the Marwari and the Magra. A similar, though short-term, study on the semen of an exotic breed, the Corriedale, was also undertaken.

Fifty-six experiments, each with 15 levels of P fertilizer in the presence of excess N and K fertilizer, were carried out on adjacent sites of the same field where the soil was maintained at the same low P status. Yields, in every experiment where there was a response, were related to level of P fertilizer by a diminishing-retums type curve, and fitted an inverse polynomial equation with a single parameter to define responsiveness. Responsiveness of many crops were similar but there were, nevertheless, considerable inter-crop differences.

Applications of P fertilizer increased the % P in the dry matter of lettuce and spinach as well as yields. They increased the % P in the Cruciferae and Chenopodiaceae without appreciably affecting yield. Conversely, theyhad little effect on the % P of leeks, onions, broad beans and French beans but increased yields.

When the optimum levels of P fertilizer were applied, % P (in the entire plant) of the different crops was negatively correlated with total dry weight per unit area and total uptake of P was related by a single curved relationship to total dry weight. In addition, the difference between the % P in the foliage and in the storage roots of the various root crops was asymptotically related to mean plant weight.

Percentage recovery of added P (100 kg/ha) by the different crops was largely determined by the total weight of dry matter. It varied from 1% when crop dry weight was 2 t/ha to 12% when it was 15 t/ha.

Applications of phosphate suppressed leaf scorch of spinach. On occasion they alleviated stem rot in summer cabbage and influenced the bolting of onions and the number of defective Brussels sprouts. Otherwise, the effects on quality were small.

Agri-silviculture is the simultaneous husbandry of forest tree crops and food crops. In the first 1–3 years of forest crops such as teak, pine and Gmelina arborea, they are interplanted with food crops such as yam, maize and cassava in plantation. In humid countries such as Nigeria and Burma agrisilviculture evolved out of the needs to boost food production and satisfy labourers who work on afforestation programmes.

A pot culture experiment with 30 Old Alluvium soils was conducted to evaluate the critical Zn concentration in soils and wheat plants for predicting the response to Zn application. The DTPA-extractable Zn (available Zn) in these soils ranged from 0·37 to 4·50 mg/kg soil, and total Zn in the wheat plants of the control pots ranged from 17 to 65 mg/kg wheat shoots. The DTPA-extractable Zn was found to be positively correlated with organic carbon (r = 0·72, P < 0·01), negatively with pH (r = –0·48, P < 0·01) and positively with wheat tissue Zn concentration (r = 0·98, P < 0·01). The Bray's ‘per cent’ yield value was positively correlated with ‘available’ Zn (r = 0·82, P < 0·01) and plant tissue Zn (r = 0·85, P < 0·01). The critical concentration of Zn in soil and plant below which responses to applied Zn may be expected was 0·65 mg Zn/kg soil and 24·5 mg Zn/kg wheat shoots respectively. All the soils below 0·65 mg Zn/kg soil responded to Zn application where the ‘per cent’ response at 5mg added Zn/kg soil ranged from 22·65 to 100·00 with a mean value of 58·24 whereas at 10 mg added Zn/kg soil, it ranged from 13·08 to 97·14 with a mean value of 40·08. Soils above the critical limit negatively responded to Zn application.

Data from experimental and observational studies, in agriculture as in many other contexts, are often unbalanced in respect of important classifications. Treatments may be unequally replicated, some combinations of factors may be omitted, animals may die for reasons unconnected with an experiment. Unless means are adjusted in some manner that eliminates disturbance from unequal representation of different categories, comparisons between treatments may be thoroughly misleading. Optimal procedures for the simpler situations have been familiar to statisticians for a long time. They have been little used by other scientists analysing their own data, in part because of the computational labour and in part because their nature has not been properly understood.

Modern computing power removes all excuse for the retention of methods that may be actively misleading because they bias summaries of data, or that are at best inefficient in their failure to estimate comparisons as precisely as is possible. The only remaining barrier is the mistaken belief that good methods are either so complicated that they can be comprehended only by professional statisticians or so devious that truth is destroyed rather than exhibited. This paper attempts to show good methods as inherently rational, to explain their main properties, and to illustrate them on examples small enough for the processes to be clear. The paper neither expounds statistical theory nor instructs in computing practice.

In ruminant animals, it is well known that the physical form of feed considerably affects rumination behaviour, which is considered to be an importantfunction related to the utilization of roughage feed (Balch, 1971; Ørskov, Fraser & Gordon, 1974; Morgan & Campling, 1978). It is also clear that the act of chewing during rumination has an important role in the reduction of particle size of ingested food, and this might facilitate microbial degradation in the rumen (Umezu, 1966). Furthermore, it is also obvious that rumination behaviour is considerably influenced by changes of dietary chemical composition. Campling, Freer & Balch (1962) observed a marked reduction in time spent ruminating when urea was added to an oat-straw diet in cows, and they suggested that there are important interactions between the two factors (chewing and bacterial degradation) responsible for reducing particle size in the reticulo-rumen. Similarly, there are several investigations of rumination behaviour in relation to the digestion of roughage diet in sheep and cattle (Welch & Smith, 1969a; Thomas, Kelly & Wait, 1976; Sudweeks et al. 1975).

The effect of level of N fertilizer on the composition, yield and quality of 21 crops was studied in experiments on adjacent sites of the same field to aid in the development of fertilizer recommendations.

Yield of each of the crops first increased and then either remained the same or declined with further increases of N fertilizer. Interpretation by means of a simple model enabled response curves to be characterized by two parameters; one representing the beneficial component of the response and the other the detrimental component. Both varied greatly from crop to crop.

The magnitude of the beneficial component of the response of most non-leguminous crops was largely determined by the potential demand of the crop for nitrogen; the exceptions were some root crops which responded less than would be expected on this basis. The adverse component was serious with root crops and those crops that are in the soil for only a short period. High levels of N increased the ratio of foliage to storage root dry weights even when total dry matter was unaffected. The changes were associated with a considerable increase in the % N in the dry matter of the roots.

When crops were grown with their optimum levels of N fertilizer a simple linear. relationship between the mean %N in the dry matter and the total weight of dry matter per unit area covered all crops. Simple relationships also existed between total dry matter of non-leguminous crops and (a) the amount of N taken up by the crop from unfertilized soil, (b) the recovery of added fertilizer by the crop and (c) the beneficial component of the response of crops harvested before October.

Percentage N in the dry matter at harvest was not a sensitive indicator of the extent to which plant growth was restricted by lack of nitrogen; a difference of 0·1% N in the plant material was associated with a 10% increase in yield.

N fertilizer levels influenced the % dry matter and the incidence of crop disorders such as rotten roots and tissue discoloration, but the effects were seldom appreciable with practicable levels of fertilizer application.

Sesame (Sesamum indicum L.) is a crop of great antiquity which is widely grown in tropical and subtropical regions of Asia, Africa, South and North America and to some extent in Russia for edible oil and for animal feed purposes. The average yield in India is only about 235 kg/ha. Even under good management practices the seed yields usually remain poor and, therefore, commercial prospects of sesame production are not good.

All the pods on pigeon-pea plants were removed after 1, 2 or 3 weeks of flowering in a pot experiment and in one field trial, in order to evaluate the compensatory ability of the crop under humid tropical conditions.

Pod removal after 1 or 2 weeks of flowering led to an immediate significantly higher vegetative development and dry-matter accumulation compared with the control. Subsequent flowering led to greater pod production and pod yield than the control with seed yield increased by 21 and 50%, respectively, in the pot experiment while pod removal after 1 week of flowering increased seed yield by 15% in the field. The removal of pods after 3 weeks of flowering was detrimental to subsequent pod development and seed yield which was decreased by 56 and 34% in the pot and field experiments, respectively.

It would seem that the loss of pods produced soon after flowering starts, before active pod filling begins, can be tolerated by pigeon peas, whereas the loss of fully elongated and actively filling pods would significantly reduce seed yield in the crop.