In general, the form and function of living plants reflect a requirement to maximize their interaction with the environment in order to harvest more effectively the energy and materials they require (Corner 1964). Thus, fluctuations in the aerial environment exert a considerable effect upon the physiology of the plant and lead to initiation of a range of responses. Changes in temperature are known to exert a pronounced effect on the growth of plants, and hence upon their productivity (Ong & Baker 1985). Most temperate species spend the majority of their life at mean temperatures below the optimum for their growth, and there are marked genetic differences between plant species in their ability to tolerate nonoptimal temperatures (Pollock & Eagles 1988). This review summarizes some of the ways in which plants are known to sense and respond to temperature change and discusses the potential for improving growth and performance at nonoptimal temperatures. Discussion concentrates upon temperate grasses and cereals because of their suitability as experimental material and because of their economic importance. Consequently, this review is largely concerned with responses to low temperatures, but some responses of tropical cereals to high temperatures are also described.

The effect of soil water supply on efficiency of N use in wheat and NO3-N accumulation in the soil profile was studied in a field experiment on loamy sand soil at Boothgarh in Hoshiarpur, Punjab, India in 1980/81. As the water supply increased from 509 to 548 mm, the response to N application increased by 11·5 kg grain/kg N and apparent N recovery by 11% at optimum N application rates. The increase in N recovery with additional water supply was related to the NO3-N use from the profile. The accumulation of NO3-N in the profile was related linearly to rate of N application but inversely to water supply up to an optimum value of 580 mm. Further increase in water supply did not increase fertilizer N use. If both the amount of moisture stored in the profile at sowing and the probable rainfall during the crop season are known, the efficiency of fertilizer N use can be predicted.

The effects of single and repeated early applications of chlormequat (at up to and including Zadoks growth stage (GS) 32) on the growth and development of the winter oat cultivar Bulwark were investigated in two seasons of field experiments at Tenby, UK. Early dry matter production, shoot initiation and subsequent shoot survival were unaffected by all treatments. At maturity, reductions in stem length and lodging were best achieved by chlormequat applied at GS 32; this treatment shortened stems by an average of 24% in 1985/86 and 31% in 1986/87. Earlier applications were significantly less effective.

Grain yields were unaffected by single and repeated early applications of chlormequat. However, in both seasons, yields were significantly reduced by repeated treatments with chlormequat, including application at GS 32. Individual grain weight was the yield component principally affected.

Two rotation field experiments of a single forage–cereal cycle each were conducted on a clay loam soil (Vertic Cambisols) in Cyprus in 1982–85. The forages in the first cycle (1982–84) were pure stands of vetch (Vicia saliva L.), peas (Pisum sativum L.) and oats (Avena sativa L.) and mixtures of each of the two legumes with oats at sowing ratios of 90:10, 75:25, 50:50 and 25:75 grown with 15, 30, 60, and 90 kg fertilizer N/ha. For the second cycle (1983–85) the same crops were used but with only two mixture ratios (legume:oats 90:10 and 75:25), and with 0, 15, 45 and 90 kg N/ha. Barley (Hordeum vulgare L.) was the test crop after the forages in both cycles. Legumes did not respond to N fertilization, whereas the yields of the pure stand of oats and the mixtures increased. The legumes yielded more N but less dry matter than the cereals. The grain and N yields of barley were higher after legumes than after oats, with intermediate yields after the mixtures. Barley responded to residual N fertilizer, irrespective of whether it was applied on legumes, oats or mixtures in the previous year. When choosing the forage to be grown, farmers should consider (a) the need for roughage and protein, (b) the costs of N fertilizer for crops and of protein for animal feed stuff and (c) the rotational role of the crop.

Plots at Hurley, UK, sown to perennial ryegrass (Lolium perenne) in 1968 were given 188, 376 or 752 kg N fertilizer/ha per year. Half of each plot at the lowest and highest rate of N fertilizer was given a severe pesticide (phorate) treatment at frequent intervals. Plots were cut and their herbage yield was assessed on five occasions during each of the following 20 years. There was no indication of a long-term decline in total annual herbage yield at any rate of N fertilizer use. The severe pesticide regime enhanced yield by an average of 14·6%/year during the first 11 years, but tended to have little effect or to reduce yields subseque7tly. Initially (1969) all treatment plots comprised at least 85% perennial ryegrass, but this proportion declined to between 67% (medium N) and 40% (high N) by 1989. Apart from ryegrass, the main sward components were species of the grasses Agrostis, Poa, and Elymus and traces of the broad-leaved species Stellaria media and Taraxacum officinale.

Crop yields and soil properties of grazed grass–clover pasture and arable plots at the Grasslands Research Institute, Hurley, UK. were measured over 30 years (1955–84). Yields of herbage declined from 9·6 t DM/ha in the first harvest year to 4–6 t in years 5 to 8 but subseque8tly increased (after liming) to give a mean yield of 8·2 t for 1963–84. Yields of barley in plots receiving a seed-bed application of 45 kg N/ha averaged (1962–80) 3·1 t/ha and, in plots top dressed with an additional 100 kg N/ha, 4·1 t/ha but after 1969, response to additional N and yield under both N treatments showed a marked decline.

Soil carbon increased during the first 10 years under pasture from 1·2 to 1·8% and soil N from 0·13 to 0·17%. These increases, equivalent to 1 t C and 75 kg N/ha per year, occurred at the same time as the initial decline in herbage yield noted above. During the subsequent 20 years, C and N remained relatively constant and herbage yields increased. Soil C and N decreased under 30 years' arable management to 0·8% C and 0·1% N, equivalent to a loss of 290 kg C and 25 kg N/ha per year.

Major factors that appear to have influenced herbage yields were (i) soil moisture, (ii) accumulation of N in the soil organic matter during the first 8 years and the subsequent increase in the rate of mineralization, and (iii) the action of the earthworm population in creating a stone-free, fertile layer of top soil. As for barley, it was concluded that yields could not be maintained indefinitely on this soil series, as a result of the loss of soil structure and the development of a compact layer of soil at plough depth.

Over two years (1984/85 and 1986/87), monthly sampling of shoots of white clover plants compared the populations of white clover in mixed swards at Palmerston North, New Zealand, under set stocking, rotational grazing and a combination of both systems, at a common stocking rate of 22·5 ewe equivalents/ha.

The frequency distributions of shoot (or stolon) dry weight per plant in each population over the study period was described by a log-normal model, which indicated that the populations consisted of many small individuals and few large individuals. Such inequality of shoot dry weight within populations is commonly termed size hierarchy; a statistic giving a measure of such size hierarchy is the Gini coefficient. The populations under different managements had similar Gini coefficients which differed little among seasons or between years. Lack of significant correlation between the Gini coefficient and mean shoot dry weight per plant of each population indicated that, in these white clover populations, size hierarchy was independent of mean plant size.

These results were considered in relation to the clonal growth of white clover in grazed swards and it is suggested that the variable nature of death of older basal stolons makes an important contribution to the variability in size of individual plants and hence to size hierarchy. As size hierarchy, as assessed by Gini coefficients, was relatively stable in these populations over 3½ years, it appears that clonal growth of white clover incorporates sufficient variability within the growth and death processes at the individual plant level to maintain the size hierarchy, irrespective of variations in mean plant size of populations.

During the growing season of 1988 at the Agricultural Research Institute, Hillsborough, cattle slurry was kept whole or was separated through screens with mesh sizes of 5 mm, 3 mm, 0·25 mm and 0·015 mm. The resulting slurry liquids were either left unacidified or were acidified to pH 5.5 with 5M-H2SO4 prior to application at 75 kg NH+4-N/ha to perennial ryegrass in the field. Volatilization of ammonia was measured using ventilated enclosures. Plots treated with a range of rates of inorganic fertilizer N were used to construct response curves to assess the efficiency of N in slurry for herbage production compared with equinitrogenous inorganic fertilizer.

On average over the season, acidification of cattle slurry decreased ammonia volatilization to < 15% of that from unacidified whole slurry. Compared with inorganic fertilizer N, the overall efficiency of NH+4-N was 96% in acidified slurries and 36% in unacidified whole slurry. Ammonia volatilization from separated slurries decreased with decreasing mesh size to 39% of that from whole slurry for a 0·015 mm mesh. The overall efficiency of NH4+-N in slurry separated through a 0·015 mm mesh was 76%. A highly significant linear relationship was found between NH3-N volatilized from all slurry treatments and N offtake in herbage at all three harvests.

Tannin-containing and tannin-free members of two pairs of near-isogenic lines of Vicia faba were studied in glasshouse experiments at Plant Breeding International Cambridge in 1989 to determine the cause of poor establishment of tannin-free lines in the field. Tests in field soils demonstrated marked differences in disease severity and growth, the tannin-containing lines being the more resistant to Fusarium culmorum and F. solani, the principal pathogens isolated. Susceptibility was not attributable to testa damage in tannin-free lines. In further tests on a range of varieties, some tanninfree and low-tannin, coloured-flowered varieties were as resistant as standard high-tannin lines. The results indicate that it should be possible to select for resistance to seedling infection and to subseque1t poor establishment among tannin-free lines and varieties.

Four soyabean cultivars were grown with two N application rates (50 and 300 kg N/ha) in the field at Hiroshima University, Japan, from June to August 1988. Cell membrane stability (CMS) by the polyethylene glycol (PEG) test, leaf water relations and nutrient concentrations in cell sap and leaf tissues were measured when the plants were 50 days old, in the uppermost fully expanded leaves.

Cell membrane stability was higher at the higher N rate, the increase over the lower rate being greater in the cultivars Lee+ and Lee–than in Tamahomare and T201. Leaf water potential was not affected by the higher rate of N application. Osmotic adjustment, which was independent of water stress, was observed with the higher rate of N and it was higher in Lee + and Lee–than in Tamahomare and T201. It is suggested that osmotic potential in leaf tissues may influence CMS measured by the PEG test. Solute concentrations in cell sap and leaf tissues were higher at the higher N rate. Sugar and K were the major contributors to osmotic potential.

The pollination requirements of the white lupin cultivar Vladimir were investigated in a glasshouse. Five treatments were compared in which the flowers of each plant were (i) left to autopollinate, (ii) shaken, (iii) tripped, (iv) self pollinated or (v) cross pollinated. Pollination treatments had no effect on the numbers of flowers, pods or seeds or weight of seed produced by the plants but did slightly shorten the duration of flowering, the effect being greatest when flowers were tripped or cross pollinated and least when they were shaken or self pollinated.

The proportion of flowers that developed into pods differed with raceme, with secondaries producing more than tertiaries > primaries > quaternaries; the number of seeds per pod differed in the same order, but mean single-seed weight decreased on successive racemes. However, pollination treatments had no effect on the distribution of flowers, pods or seeds on the plant.

This crop species is clearly highly self fertile and self pollinating and likely to yield well without insect pollination.

In studies of carrots sown on three dates at Wellesbourne in 1986, mean time to seedling emergence and spread of times to seedling emergence of untreated and fluid-drilled seeds increased as seed-bed moisture at sowing decreased. These differences were not observed with irrigation before sowing.

Osmotic priming increased the percentage of seeds with emerged radicles at the time of fluid drilling from 17% in the untreated control to 56%. Irrespective of seed-bed moisture, time to emergence was shorter from primed germinating seeds than from germinating seeds, both treatments giving earlier seedling emergence than untreated seeds. Seedling shoot weight was greater from treated than from untreated seeds.

Seed-bed characteristics on unirrigated plots had no effect on seedling emergence when soil moisture was adequate but, where soil moisture was limiting, rolling the seed bed to increase capillarity resulted in 79% emergence compared with the 67% average from seed beds that were not rolled. Application of a soil conditioner to stabilize the seed-bed surface structure generally improved emergence when rain fell soon after sowing. The results suggested that a combination of seed-bed and seed treatments can significantly improve the predictability of crop establishment of carrots on different dates.

Data from 62 field experiments on various crops in 1973–85 at Rothamsted Experimental Station were used to obtain knowledge about the heterogeneity of variance of yield. The results suggest that explaining heterogeneity of variance in terms of a mean–variance relationship is inadequate; better understanding of the patterns of changing variance can be gained by investigating the influence of the experimental factors on the variance. The effects of the factor ‘amount of nitrogen’ on mean yield and on the variance were positively correlated; factors which tend to affect variability but not the mean were identified. Suggestions are given for controlling variance but the need for further experimental evidence is emphasized.

In two consecutive lactations in 1985/86 and 1986/87, 60 cows were injected with 500 mg of Sometribove, methionyl bovine somatotropin (BST), or a placebo at day 60±3 post partum; injections were repeated fortnightly for the remainder of each lactation. All cows received the same complete diet (metabolizable energy (ME) 11·5 MJ/kg dry matter; crude protein 169 g/kg dry matter) ad libitum for the first 20 weeks of lactation, after which, energy concentration of the diet was reduced as yield declined.

The administration of BST significantly increased milk yield by 4·3 (P < 0·001) and 3·1 kg/day (P < 0·05) in the first and second lactations, respectively. The mean dry matter (DM) intake of 16·3 kg/day in lactation 1 for the treated group was significantly (P < 0·01) higher than the 14·7 kg/day recorded for the control group. The response in intake was not significant until 6 weeks after treatment started. However, in the pretreatment period of the subseque1t lactation, the DM intake of the cows in the treated group was already significantly (P < 0·01) higher (+ 1·2 kg/day) than that of the control group. Although the DM intake in the treatment period of lactation 2 was also higher for the treated group, the difference from the control group was not significant.

With the exception of a significant (P < 0·01) increase in milk protein (control 35·3 g/kg v. treated 36·3 g/kg) in lactation 1, overall milk composition was unaffected by BST treatment. The administration of BST increased apparent feed efficiency (kg 4% fat-corrected milk (FCM)/kg DM intake) by 6 and 10% and gross feed efficiency (kg FCM/MJ ME) by 8 and 6% in lactations 1 and 2, respectively.

Animals treated with BST had significantly (P < 0·05) lower scores for body condition during the treatment period in lactation 1. This was, however, recouped during the dry period so that both groups had similar body conditions before the start of the next treatment period.

BST administered every 14 days in a prolonged-release formulation increased the yields of milk and milk constituents and apparent and gross feed efficiency during two consecutive lactations of treatment.

There were 45 cows in each of the control and treated groups in the first lactation of the study and, of these, 30 cows in each group received the same treatment for a second lactation. The study was conducted during the winters of 1985/86 and 1986/87. Injections of either bovine somatotropin (BST) or a placebo were started at day 60±3 post partum and repeated at fortnightly intervals for the remainder of each lactation.

Body temperature and heart and respiration rates were unaffected by the administration of BST. Minor abnormalities were noted in individual animals in both groups, hygromata of the tarsus or carpus being most common. However, no health problems could be related to treatment. There was no evidence that the administration of BST increased the incidence of metabolic disorders. Ketosis was not observed in either lactation.

Manufacturing properties and quality of Cheddar, Wensleydale and Cheshire cheeses made either with milk from cows treated with Sometribove, methionyl bovine somatotropin (BST), or with milk from untreated control cows were compared.

The source of milk did not affect manufacturing properties in terms of processing time. Renneting times for Cheddar cheese made from control milk and that made from milk produced by cows treated with BST were 57 and 55 min, respectively. Corresponding values for Wensleydale and Cheshire cheeses were 50 and 53 min and 49 and 46 min, respectively. The source of milk did not influence cheese composition in terms of moisture, fat, protein and salt. When cheeses were submitted to the National Grading Service, those made from milk produced by BST-treated cows were indistinguishable from those made from control milk. The total points scored (maximum 100) for Cheddar, Wensleydale and Cheshire cheeses made with milk from control and BST-treated cows were 89 and 88, 85 and 85, and 86 and 86, respectively.

Estimates of dry matter intake reported by Barlow el al. in this journal in 1988 appeared low for grazing beef cows during early lactation. Further estimates for 144 crossbred and Hereford cows at Grafton, NSW, support the contention that the previous estimates for early lactation were biased downward. This appears to have been simply a scale effect as rankings of pastures and cow genotypes remained unchanged, although the magnitude of differences did alter.

Six Brahman (Bos indicus) steers implanted with 300 mg trenbolone acetate and six similar nonimplanted steers were fed a low-quality, low-protein roughage diet at 10 g dry matter/kg live weight. They were housed in individual pens for 27 days in 1987 before being placed in metabolism crates for measurement of digestibility, nitrogen retention, and the rate of protein synthesis in the whole body and fractional rate in three muscles and hide.

The rate of weight loss of steers treated with trenbolone acetate was significantly (P < 0·05) less than that of controls (–0·34 v. –0·56 kg/day). Treated steers excreted significantly (P < 0·05) less 3-methylhistidine (128 v. 202 μmol) and urea N (0·5 v. 1·6 g/day) in urine. There was a more favourable nitrogen retention in animals implanted with trenbolone acetate (–10 v. –12g/day) but this difference was not significant.

Whole-body protein synthesis, calculated from plasma leucine flux, was 511 and 508 g/day in treated and control steers, respectively. Treatment also had no significant effect on the fractional rate of protein synthesis, in skeletal muscle or in hide. The fractional rate of synthesis in muscle was < 1%/day and in hide < 2%/day. These results are discussed in relation to the protein and energy metabolism of undernourished cattle.

Silage effluents with pH values of 4·1, 4·7 and 5·1 (normal (N), high (H) and very high (VH), respectively) were treated with hydrochloric acid to bring the pH of each effluent to 40. The effluents were offered to finishing pigs in a randomized block experiment (November 1983 to February 1984 at the Agricultural Research Institute of Northern Ireland) with dietary treatments which comprised a control diet in which meal was given to a standard scale of feeding, or effluent diets N, H and VH, in which the effluent was substituted on a dry matter basis for 150g/kg of the standard scale of feeding. In a negative control diet, meal was offered at 850 g/kg of the control diet. Water or effluent as appropriate was given on the meal, and water was also available ad libitum. A total of 60 pigs was used with mean initial and final live weights (LW) of 37·1 and 86·4 kg, respectively.

Effluents H and VH had higher ammonia N: total N and lower amino acid N: total N and lactic acid contents than effluent N. The rates of LW gain, feed conversion ratios, backfat thicknesses and the proportions of lean and fat in the ham joint were similar for the control diet and effluent diets N and H, indicating a similar nutritive value for all three. Effluent diet VH gave a significantly (P < 0·01) reduced rate of LW gain and poorer feed conversion ratio than diets N and H. While diet VH was consumed reluctantly with widespread refusals (mean refusal 28 g DM/kg DM offered), the feed refusals together with the lower lysine and threonine contents of the effluent would not have produced the observed degree of deterioration in growth and feed conversion. Possible reasons for the lower nutritive value are undetectable wastage of feed or the presence of an antinutritional factor in effluent VHG, or a combination of both factors.

Water and energy metabolism were measured in six growing female New Zealand White rabbits placed individually in metabolism cages within chambers at 20 or 30 °C and offered a diet high (H) or low (L) in roughage. For water balance measurements, feed intake was higher at 20 °C than at 30 °C; mean water balance of rabbits on diet L was 7 g/day at 20 °C and 36 g/day at 30 °C. Total daily water consumption was 84–96 g/kg body weight (W), and the ratio of free water to feed intake was 1·6:1 at 20 °C and 2·3:1 at 30 °C.

Calorimetric measurements showed no differences in energy metabolism of rabbits due to diet or temperature. Metabolizable energy needed for maintenance was 357 kJ/kg W0·75 per day and was used with an efficiency of 0·44 for weight gain.