Two factor designs are considered in which the levels of the two factors are linked (e.g. certain diallel cross and reciprocal transplant experiments). A common-row-and-column (CRAC) effect is defined as a systematic effect in the response when the treatment combination has the same level of the two factors. The efficiency of designs for estimating CRAC effects is discussed. Square designs with various levels of extra replication for CRAC treatment combinations and strand designs, based on omitting certain treatment combinations for which the two factors do not have the same level of the two factors, are defined and their efficiency examined.

In the design for optimal estimation of a single CRAC effect in square or strand designs, half the total replication is in CRAC treatment combinations and half in the other combinations. This gives a major increase in design efficiency, particularly where the number of levels of the factors is five or greater. Designs with equal replication of all treatment combinations are generally very inefficient. The optimal square and strand designs have equal efficiency for estimating the CRAC effect, but the strand designs are more flexible in practice, requiring fewer plots to make a complete block and this relative flexibility increases with the number of levels of the factors.

Most of the square or strand designs allow the estimation of n separate CRAC effects where n is the number of levels of the factors. Optimal square designs for estimation of separate CRAC effects give replication of CRAC combinations as ((n–1)(n–2)/2)½ times that of other combinations, which is different from the optimum value for estimating the single CRAC effect. However, for many designs, the difference is not of major importance because optimal designs for the estimation of a single CRAC effect are relatively insensitive to small deviations from the optimal ratio.

When the replication of CRAC combinations differs from that of the other treatment combinations, the designs are less efficient in estimating the factor effects. Where estimation of factor effects is also of relevance, a trade-off between increased efficiency in the estimation of CRAC effects and of factor effects may be desirable.

Strand designs are not as efficient as the optimal square design for estimating separate CRAC effects. For three-strand designs of even order examined, the CRAC effects are confounded with factor effects and are not all separately estimable. When factor effects and/or separate CRAC effects are of interest, square are preferable to strand designs.

Some of the results generalize readily to three linked factors. There is a three-factor CRAC effect and three two-factor CRAC effects. The replication allocation for CRAC relative to other combinations, to optimize the estimation of the three-factor CRAC effect is (n–1)(n–2)/2 for the three-factor CRAC combinations and (n–2)/2 for the two-factor CRAC combinations. Selection to optimize the estimation of the two-factor CRAC effects gives very different optimal allocation.

Methods are described for estimating interplot interference in variety trials. The effects on variety yield performance of interference from varieties in adjacent plots was studied for more than 600 UK National and Recommended List variety trials in the years 1987–1991. The crops examined were barley, wheat, oats, grain peas, beans, oilseed rape, forage maize and sugarbeet. Estimates of differences in variety performance may be biased by as much as 4% in some crops. Factors associated with interference included plant height, standing ability and disease susceptibility. An approach is outlined for limiting the effects of interference through controlled randomization within trials.

Omitting or changing various inputs to winter wheat, cv. Norman, were studied over the 1985/86, 1986/87 and 1987/88 crop years in Northern Ireland. Control plots received chemical inputs as commonly applied to intensively managed wheat, including herbicide, two broad spectrum fungicides and chlormequat and yielded 7·5, 9·3 and 6·3 t/ha in 1986, 1987 and 1988 respectively. Other treatments included no herbicide, no chlormequat, no late fungicide, no fungicide, no sprays, half seed rate, late sowing and late sowing with no sprays. Later sowing decreased take-all in 1986 and increased yield. All other treatments either reduced yield or had no effect. Omitting fungicide gave consistent and large decreases in yield by depressing the 1000-grain weight but the importance of early and late applications varied between years. The yield reductions from omitting fungicide were generally larger than those reported elsewhere. Lodging occurred in all years but chlormequat reduced it only in 1988 and gave a yield increase only in 1986, although the number of grains/ear was increased in all years. There was no evidence of any interactions between different pesticide inputs or with seed rate or sowing date.

The possible interacting effects of shading and N supply on number of grains of Triticum aestivum L. (cv. Buck Ñandú) were investigated at Balcarce, Argentina, during the 1988/89 and 1989/90 growing seasons. Shading was imposed from c. 13 days before anthesis to 6 days after, and four rates of N fertilization were supplied within each shading treatment around the date of terminal spikelet formation. Water and other nutrients were not limiting.

Total grain yield was strongly correlated with grain number/m2, regardless of shading or N supply. At the highest N rates, grain number and dry weight of spikes at anthesis were linearly related to a photothermal quotient, i.e. the ratio of intercepted photosynthetically active radiation (PAR) to mean temperature minus 4·5 °C, during the period from 20 days before anthesis to 10 days after. The response of grain number to the photothermal quotient was interpreted in terms of the supply of assimilates to the spike at anthesis, which determined flower survival. The response of dry weight of spikes to photothermal quotient was interpreted in terms of crop growth rate since there was a linear relationship between crop growth rate and intercepted radiation. The lowest N rates reduced the number of grains/m2, at any given photothermal quotient. Since the reduction in grain number also occurred at any given dry weight of spikes, it cannot be explained by a reduced supply of assimilates to the spikes. Grain number responded directly to the supply of N to the spike, probably through the survival of differentiated flowers. The relationship between spike growth rate and crop growth rate was not affected by N supply. Crop growth rate was reduced by reduced N supply, because less radiation was intercepted and because radiation-use efficiency was lowered. These results indicate that current models for determining yield and number of grains/m2, based on crop growth, are not adequate when N is deficient.

Field experiments were undertaken at the University of Wales, Aberystwyth over two seasons to investigate the effects of N fertilizer rate and timing on the development of Septoria diseases (principally Septoria tritici) in winter wheat and its subsequent effect on crop growth. Rates of 100, 200 and 300 kg/ha N were applied at GS23, GS30, or split (GS23/30/38) in 1985/86 and 100 and 250 kg/ha N at GS23, GS30, GS23/30, or GS23/30/37 in 1987/88. Foliar diseases were allowed to develop or were controlled with fungicide. Early disease levels were temporarily enhanced soon after N application and were greater at the higher rates of N. In April and June a significant correlation between disease infection and concentration of N in leaf tissue was shown. Despite differences in the level of early infection, dry matter production up to and including anthesis was unaffected by disease. Infection during grain filling caused premature senescence of the flag leaf which significantly reduced grain yield, primarily through a reduction in grain size. Loss of yield was directly related to severity of disease on the flag leaf in early July, with a yield loss of 0·45% for every 1% increase in disease infection. The severity of Septoria post-anthesis increased with increasing rate of fertilizer N by an average of 11·1% per 100 kg/ha N (translating to a potential yield loss of 5% for each 100 kg/ha N given as fertilizer), but was unaffected by its timing of application.

A simple model of vernalization, originally developed to quantify the vernalization response of fieldgrown carrots, was fitted to previously published experimental results for winter wheat cv. Norin 27. The optimum temperature for vernalization indicated by the model was c. 5·2 °C, as this induced the fastest progress to final leaf appearance, expressed as the reciprocal of number of days from sowing to final leaf. This rate decreased linearly with temperature rise or fall on either side of the optimum, extrapolating to zero at –4·8 °C (Tmin) and 26·6 °C (Tmax). When all the treatment temperatures and durations were expressed as vernalizing degree days > –4·8 °C (V °C d), there was a linear increase in post-treatment development rate with increasing vernalization up to c. 275 V °C d. Ending the effective treatment duration for vernalization at the estimated time of initiation of the final leaf primordia brought many of the data points closer to the linear trend which described the rest of the data.

Effects of using leaf number, which is linearly related to thermal time, instead of days as the unit of time to compensate for temperature differences in the original experiment were examined. Unvernalized plants had the potential to produce 18 leaves before flowering and therefore rates were expressed as the fraction of the potential total leaf number that each new leaf represented. All plants were assumed to have an initial development rate of 1/18 per leaf. This rate was assumed to increase linearly with time during the vernalizing treatment periods and then remain constant after treatment until the final leaf appeared. Leaf numbers reported from the original experiment were used with these assumptions to estimate the rate at the end of each treatment. The relationship between these rates and treatment temperatures was similar to that for rates based on post-treatment durations. There was an optimum temperature c. 5·5 °C and Tmin and Tmax of –5·1 and 18·8 °C estimated by extrapolating the decreasing linear trends to the base rate of 1/18. When plotted against V °C d calculated from these temperatures, the rates from the full data set were well represented by the model line which had been fitted to the data from just one treatment duration.

The rate of leaf appearance of barley varies substantially with time of sowing. This variation has been related to both the length and the rate of change of photoperiod at the time of plant emergence. An outdoor pot experiment was conducted to test if rate of change of photoperiod directly affects phasic development and rate of leaf emergence of spring barley. Two photoperiod-sensitive cultivars (Bandulla and Galleon) were subjected to five photoperiod regimes: two constant photoperiods, of 14 and 15·5 h, and three different rates of change of photoperiod of c. 2, 9 and 13 min/day from seedling emergence to awn initiation.

Photoperiod treatments significantly affected the duration from seedling emergence to awn initiation in both cultivars. Rate of change of photoperiod did not affect the rate of development towards awn initiation independently of the absolute daylength it produced. Although Bandulla had a longer duration than Galleon at any photoperiod regime, the cultivars did not vary in their sensitivity to photoperiod. When this phase was divided into the leaf initiation (LI) and spikelet initiation (SI) phases, it was evident that the sensitivity to photoperiod was not constant, being in general higher during the SI than during the LI phase. However, the magnitude of the change in sensitivity was cultivar-dependent, indicating that sensitivity to photoperiod during the different phases could be under independent genetic control.

Final numbers of primordia (leaves together with maximum spikelet number) were negatively affected by increasing photoperiods, but once again, there was no evidence of any effect of the rate of change of photoperiod which was independent of the average photoperiod. Both cultivars showed similar sensitivities for final leaf number but maximum spikelet number was more sensitive to photoperiod in Galleon than in Bandulla.

Highly significant linear relationships between leaf number and thermal time were found for all combinations of cultivars and photoperiod regimes (r2 > 0·98). The rate of leaf appearance (RLA) was similar for both cultivars (c. 0·0185 leaves/°Cd) and did not alter during plant development or in response to the change in photoperiod at awn initiation. The range in RLA was greater for Galleon (0·0170–0·0205 leaves/°Cd) than for Bandulla (0·0173–0·0186 leaves/°Cd). Neither of these cultivars exhibited a significant relationship between rate of leaf emergence and photoperiod or rate of change of photoperiod. The lack of significant relationships between RLA and length or rate of change of photoperiod is in contrast with previous reports using time of sowing as a main treatment.

Trials were established at two sites in 1982 to compare six 2-course rotations, in which barley alternated with feed legumes (pure or mixed with barley), fallow or barley, in partial factorial combination with different N and P fertilizer regimes. This paper summarizes data from the second to the seventh cropping seasons.

Barley yielded most after fallow and more after legumes than after barley; but, on the basis of total dry matter production over both courses, barley-legume rotations outyielded barley-fallow and barley-barley rotations. In terms of net nitrogen offtake, which may be taken as a rough measure of feed value, barley-legumes surpassed the other rotations by c. 100% at both sites, whether the comparison was made on values obtained from plots with or without added fertilizer. Of the legumes, Lathyrus sativus was slightly more productive than Vicia saliva, and this carried over into the rotational yield of both dry matter and nitrogen. Rotations involving pure legumes were slightly more productive than those involving legume/barley mixtures.

Yields of all crops and rotations varied greatly from year to year depending on rainfall. Those of the barley-fallow rotation were much less variable than those of the barley-barley rotation at the wetter site (323±103 mm), with the barley-legume rotation intermediate; but variability was much greater at the drier site (250±79 mm), with much less difference between rotations. Nevertheless, fertilizer use and rotations including legumes gave a significant yield advantage under the driest conditions.

Tiller weight, lamina length, lamina growth rate, appearance interval and lamina utilization were examined in Lolium perenne, Agrostis stolonifera, Holcus lanatus and Poa trivialis. Marked tillers were measured in situ over 14–21 day periods in a continuously grazed permanent pasture under steady state management in Devon, UK, in 1985, 1986 and 1987 on plots receiving either zero (0N) or 400 kg nitrogen (400N) fertilizer/ha per annum.

L. perenne was incapable of reducing its individual tiller weight or lamina length to the same extent as in the other three species during the grazing season. Lamina appearance interval was longer in L. perenne than in the other three species in most of the observation periods in both the 400N and 0N plots. Overall the lamina appearance intervals were similar between A. stolonifera, H. lanatus and P. trivialis. Lamina extension rates were greater in L. perenne than in either A. stolonifera or P. trivialis in most observation periods in both plots. H. lanatus had lower lamina extension rates than L. perenne in c. 50% of the observation periods in both plots, at other times the extension rates of the two species were similar, with one exception in early spring in the 400N plot when H. lanatus had a higher extension rate. There was no clear pattern in seasonality as to when H. lanatus had a lower extension rate than L. perenne. H. lanatusachieved either similar or higher lamina extension rates than either A. stoloniferaor P. trivialis in c. 50% of the observation periods, respectively; the periods when higher rates were observed in H. lanatus in the 400N plot occurred mainly in the spring and early summer. A. stolonifera and P. trivialis achieved similar lamina extension rates in most observation periods in both the ON and 400N plots. The specific differences observed in the lamina extension rates were emphasised when converted to growth rates by multiplying the length increments by the average weight per unit length of the expanded lamina for each species.

Lamina utilization, in terms of percentage of lamina length removed by grazing, was similar between the four grass species in most observation periods in the ON plot. A. stolonifera lost more lamina length than either L. perenne or P. trivialis in May and June and H. lanatus lost a greater percentage of lamina length than either L. perenne or P. trivialis in May in the ON plot. In the 400N plot L. perenne lost a greater percentage of lamina length than any of the other species in June and more than P. trivialis in May. A. stolonifera lost a greater percentage of lamina length than any of the other species in July, and more than either L. perenne or P. trivialis in May. This greater severity of grazing in A. stolonifera was associated with a high incidence of tillers being grazed to stubble. H. lanatus lost more lamina length than L. perenne in May. When the loss of lamina length through grazing was converted to losses in terms of weight of tissue removed, then L. perenne consistently lost more than either A. stolonifera or P. trivialis in both the 400N and 0N plots.

The implications of these differences in tissue production and utilization on competitive interactions between L. perenne and the other three species are discussed.

Thirty-five populations of Medicago polymorpha were collected from throughout Sardinia (Italy) in 1989 with a view to developing pasture species suitable for improving degraded grasslands in the northern Mediterranean basin. Herbage and seed production were compared with the Australian cultivar, Circle Valley, over 2 years at Bonassai, north Sardinia. Regeneration in the 2 years after sowing and seed size were also estimated. All variables were related to collection site parameters using multivariate analysis.

Herbage production varied between 2 and 8 t dry matter/ha, and up to 1·5 t/ha of seed was produced. K-means clustering of agronomic and morphological variables indicated that there were two groups present; one similar to variety polymorpha and one to variety vulgaris. There were two other single entry clusters, one of which contained cv. Circle Valley. Principal component analysis of the environmental variables indicated that cluster 1 (vulgaris) was more likely to come from mountainous areas where winter temperatures are low, and cluster 2 (polymorpha) from coastal areas where temperatures are mild. Regeneration of cluster 1 was better than that of cluster 2, which in turn was better than Circle Valley, indicating that populations in cluster 1 are better adapted to the management system imposed at Bonassai.

The results indicate that M. polymorpha has considerable potential to improve the grasslands of Sardinia. However, it is unlikely that imported cultivars will be successful, and it seems important that the selection of local populations should continue. Commercial seed production in Sardinia is likely to be a problem, and grazing management under the conditions of communal ownership may have to be reviewed. It is important that future research and development involves farmers and other industry groups.

The growth of sugarbeet (Beta vulgaris L.) seedlings was investigated under laboratory conditions in 1990 and 1991. Three phases were identified during heterotrophic growth in the dark, based on the changes in dry weight, carbon and nitrogen contents. The perispermic reserves were first transferred to the young seedling. All the components of the seedling (cotyledons, hypocotyl, radicle) began to elongate and their dry weights increased. The seedling separated from the seed at the end of this first period, 4 days after imbibition had started at 20 °C. During the second period, there was a major redistribution of material between organs, mainly from cotyledons to the hypocotyl. The different organs elongated rapidly. From the start of the third period, 6 days after sowing at 20 °C, the organs began to lose weight and gradually ceased to elongate. The dry weights of the whole seedlings and of each organ were proportional to the initial seed weights. The elongations and changes in dry weights obtained at different temperatures coincided when time was expressed as thermal time using a base temperature of 3·5 °C. The growth of the seedling after illumination depended on the heterotrophic growth duration. If emergence occurred after 100 °Cd, the relative growth rate of the seedling decreased. In field conditions, plant variability at the end of sugarbeet establishment is due not only to variations in seed weights but also to environmental conditions that control emergence delay.

The nutritional potential of a number of raw tropical seeds was assessed in a series of feeding trials with rats. Seed lectin reactivity was also monitored, α-amylase and trypsin inhibitory activities were determined in some of the seeds.

Abelmosclius esculentus, Chenopodium quinoa, Delonix regia, Macroptilium lathyroides, Papaver sonmiferum, Parkia biglandulosa, Sesbania arabica, Terminalia catappa, Vigna subterranea, Vigna umbellata and Vigna unguiculata seeds supported moderate rat growth. The seeds contained only low levels of essentially non-toxic lectin, moderate amounts of trypsin inhibitors and negligible quantities of a-amylase inhibitors and they have great potential as dietary protein sources for man and animals.

Artocarpus altilis, Canavalia ensiformis, Canavalia maritima, Dioclea grandiflora, Phaseolus acutifolius, Phaseolus coccineus and Phaseolus vulgaris cv. Processor, cv. Rosinha G2 and cv. Carioca 80 seeds were toxic. These seeds contained high levels of potentially toxic lectins. Other antinutritional factors may also have contributed to the high oral toxicity of some of these seeds.

Albizia adinocephala, Albizia lebbeck, Bauhinia violacea, Cassia nodosa, Cassia tora, Dioclea sclerocarpa, Entada phaseoloides, Enterolobium cyclocarpum, Leucaena leucocephala and Moringa oleifera seeds were also highly toxic but had only low levels of essentially non-toxic lectins suggesting that the toxicity was due to other anti-nutritional factors.

Bauhinia reticulata, Macrotyloma uniflorum and Tamarindus indica proteins were poorly digested and utilized. The seeds contained low levels of lectins which agglutinated only rat and cattle erythrocytes which had been pre-treated with suitable proteases. Brownea macrophylla had a similar lectin reactivity.

Ammonia volatilization from stored slurry or surface-applied slurry in the field is strongly affected by pH. Thus a simple, iterative model was developed to predict pH changes in slurry. Concentrations of NH4+/NH3, CO2/HCO3-/CO32- and volatile fatty acids were input data to the model. The model was validated by titrating 17 slurry samples collected from four cattle farms, seven pig farms and three biogas plants. Predictions of pH agreed well for 14 slurries with titration data in the pH interval from 4 to 10. Simulations indicated that microbial degradation of VFA to methane and carbon dioxide resulted in a pH increase if the carbon dioxide produced was lost to the atmosphere. There was little change in pH if the produced carbon dioxide dissolved in the slurry.

The feeding behaviour of indigenous goats grazing on natural vegetation in a semi-arid region of Zimbabwe was studied over five seasons covering a period of one year (March 1989 to March 1990). The goats were randomly allocated to two groups, A and B, which grazed for 8 h (08.00–16.00 h) and 5 h (11.00–16.00 h) daily, respectively. These grazing times were based on the current practices of smallholder farmers. The botanical and chemical composition of the diet was determined by analysing samples collected from oesophageal fistulated goats. Daily feed intake was estimated by the total faecal collection method. Feed selection varied markedly with season and was similar in both groups. From January to March (wet season) grass formed 55–70% of the forage consumed, while from May to early November (dry period) browse, herbs and fruits formed 77–91% of the diet. Dietary crude protein levels ranged from 128 to 157 g/kg over the seasons. Neutral detergent fibre content ranged from 590 to 790 g/kg, and was highest in wet months and lowest in dry months. Feed intake ranged from 33 to 60 g/kg BW0·75 or 16–28 g/kg of body weight. It appeared that a feeding duration of 5 h per day was adequate for maintenance but insufficient for optimum production.

The effects of synchronizing dietary energy and nitrogen supply in diets with a similar carbohydrate composition on microbial protein synthesis and rumen fermentation were examined in sheep. Two diets were formulated to be either synchronous (diet S) or asynchronous (diet A) for the hourly release of nitrogen (N) and energy to the rumen. Diet S contained (g/kg) 425 g wheat straw, 400 g winter barley, 150 grapeseed meal and 25 g minerals/vitamins and diet A contained 505 g wheat straw, 458.5 g winter barley, 11·5 g urea and 25 g minerals/vitamins. Both diets were fed at the rate of 1 kg/day in four equal portions, to four cannulated sheep, in two periods in a change-over design. Rumen ammonia concentrations followed the predicted hourly trend in N degradation with a peak 1 h after feeding of 10 mM for diet S and 16 mM for diet A before falling within 3 h of feeding to 4 ITIM in animals fed either diet. Rumen volatile fatty acid (VFA) concentrations followed the cyclical trend predicted by stoichiometric equations, whilst rumen VFA ratios were more stable than predicted in animals fed either diet. The observed content of rumen degradable protein and organic matter truly degraded in the rumen was similar for both diets. The increase in total CHO digested in the rumen observed with diet A (427 g/kg DM) compared with diet S (364 g/kg DM) can be attributed to the greater content of starch in the asynchronous diet, which had a high degradability. The efficiency of microbial protein synthesis (g N/kg OM truly degraded in the rumen) was 11–20% greater in animals fed the synchronous diet (S) than the asynchronous diet (A). It is concluded that microbial N production was more efficient when dietary energy and N supply were synchronized.