Thirty-five buffalo heifers were tested daily with fertile males to ascertain the age and live weight at first possible oestrus and conception. Patterns of sexual behaviour were recorded while rectal palpation was carried out to define the conditions of the ovaries and to diagnose pregnancies. The mammary gland development, as well as the intensity of lactation, were noted post partum.

1. Pubertal matings were allowed with less certainty than adult matings. The signs of oestrus were intensified by the recurrence of heat and association with the male. Homosexuality was only observed in the first and second oestrus.

2. The average age of first oestrus, first conception and first calving were 406, 647 and 963 days respectively. The body weight at first oestrus and first conception were 198 and 319 kg. respectively.

3. The number of services/conception ranged from 1 to 7 with an average of 4·25. The number of silent heats/female ranged from 1 to 4 with an average of 1·65. The period elapsing from first oestrus to first conception ranged from 52 to 438 days. Before conception, there was a period of anoestrus which ranged from 115 to 314 days, this was probably due to weak oestrus symptoms.

4. The live weights at 28, 84, 140, 196, 252 and 308 days were correlated with the age and live weight at first oestrus as well as the live weight at first conception.

5. All the buffalo-cows except two which showed oestrus conceived. Fourteen animals calved normally while fourteen aborted after 131–318 days. The gestation period ranged from 312 to 321 days with an average of 316 days. The birth weight of young ranged from 33 to 40 kg. Seven animals were not lactating while seven gave 1–2 kg. of milk.

6. Puberty phenomenon is a gradual phenomenon and is attained in steps: sexual desire, ovulation, oestrus, conception, pregnancy then lactation.

1. The mechanical analysis of a number of soils, clays, and other granular materials has been carried out by the writer's method and the results have been set out in the form of curves showing the relationship between summation percentages and logarithm of settling velocity.

2. The curves obtained are smooth. This suggests that all the necessary information as to the mechanical composition of a material can be obtained by using a relatively small number of experimental points and drawing the appropriate curve. Results obtained on one scale can readily be transferred to another scale by interpolation.

3. The most common type of curve is a sigmoid. The steepest portion of the curve represents the fraction present in greatest frequency. This has been termed by the writer the modal fraction. Two sub-types were found. In the case of certain heavy clays, the modal fraction appears to be at the lower end of the curve, whilst in the case of materials mechanically disintegrated the modal fraction is at the upper limit of particle size.

4. Composite curves are obtained in certain cases. They may represent mixtures of different materials, or soils in the earlier stages of formation from rock.

5. The lower limit of particle size appears to be represented by log in the case of normal soils and clays. In the case of certain finely divided but non-plastic materials, the lower limit appears to be in the region of . It is suggested that a better picture of the properties of soils would be obtained by taking the limit for clay at this poiDt instead of at log .

6. The use of the logarithm of settling velocity as a measure of particle size is discussed and it is suggested that the use of negative logarithms might be avoided by using log (v × 107), for which the symbol Lv is proposed.

Forty Large White × Wessex pigs were fed from 7 to 27 kg live weight, on diets containing either 28 or 14% crude protein. From 27 to 113 kg live weight the pigs were fed a common diet ad libitum. Pigs from each treatment were killed at 27, 54, 82 and 113 kg live weight and the chemical composition was determined. From 7 to 27 kg and from 27 to 113 kg live weight, daily gain, food intake and food conversion efficiency were similar for the two treatments.

At 27 kg, pigs on the lower protein diet were fatter and contained less lean tissue. Between 27 and 54 kg live weight the pigs previously on the higher protein diet deposited a much greater amount of fat than the pigs on the lower protein diet, so that by 54 kg the pigs initially on the lower protein contained less fat. At 82 and 113 kg live weight the pigs given the lower protein diet initially also tended to contain less fat. These findings are discussed with particular reference to the voluntary energy intake of growing pigs.

The present investigation is the sequel to an earlier pig-feeding trial in which a diet supplying only half the standard amounts of protein supplement, in the form of a mixture of feeding blood meal, dried separated milk and ex. soya bean meal, was found to give as good results from the standpoints of rate of growth, efficiency of food conversion and carcass leanness and quality, as a diet that contained the full standard allowances.

It was considered desirable to repeat this initial exploratory experiment, making use this time of a totally different form of protein supplement, since the question of the biological efficiency of the protein supply might conceivably be of critical importance at the lower levels of protein feeding. In the present trial, therefore, the mixture of the three protein foods was replaced by a single protein food, namely, white fish meal. The standard-protein treatment C supplied 12% of white fish meal up to 90 lb. live-weight, 10% from 90 to 150 lb., whilst from 150 lb. to slaughter at about 200 lb. live-weight, the white fish meal was omitted from the diet, its place being taken by 5% of ex. soya bean meal. The corresponding allowances of protein food in the low-protein treatment B were 5, 4 and 3% respectively. The low-protein treatment A, which was introduced as a link between the previous and the present trials, contained the protein supplement composed of feeding blood meal, dried separated milk and ex. soya bean meal and supplied the same amount of digestible protein as was contained in the low-protein treatment B.

Three plastic formulations of both Dursban and Dimilin were tested as controlledrelease pellets against larvae of C. pipiens and A. aegypti. Each pellet had an initial concentration of 0·116 μg/l. Results showed that Dimilin did not appear to give as high percentages of larval mortality as Dursban although in most cases a clearly delayed inhibition of adult emergence was noted. Larval mortalities of 90–100% were obtained in 3–14 (in the case of C pipiens) and 1–12 (in the case of A. aegypti) days when the second instar larvae were treated with Dursban formulations. With the Dimilin formulations no more than 31% (in case of C. pipiens) and 24% (in case of A. aegypti) larval mortalities occurred up to 13–14 days until completion of pupation.

The results obtained showed that formulation 1 (polyvinyl chloride + dibutyl phthalate + toxicant) was the most effective of the three tested ones, followed by formulations 2 (polyvinyl chloride + dibutyl phthalate + triton X·100 + toxicant) and 3 (polyvinyl chloride + dibutyl phthalate + triton X·100 + ammonium carbonate + citric acid + water + toxicant), respectively. This was so on the basis both of percentage of larval mortality and time post-treatment. The durations of effectiveness of Dursban pellets were 123, 94 and 86 days for the formulations 1, 2 and 3 respectively. However, on the basis of number of tests with over 90% larval mortality, formulation 2 gave the highest level (13 tests) followed by formulation 3 (12 tests) and finally formulation 1 (10 tests). Therefore, consideration must be given to the cumulative effect of these factors of activity when using these formulations for control of mosquito larvae. Activity indices also proved that there was no definite variation in the toxicity of both DDT-susceptible and resistant strains against these formulations. Such results were obtained with Dimilin which confirmed that formulation 3 was the most effective one. The durations of effectiveness were about 112–115 (11–67% inhibition of adult emergence), 111–113 (11–93% inhibition of adult emergence) and 102–113 (12–100% inhibition of adult emergence) days for formulations 1, 2 and 3, respectively, for the susceptible and resistant lines of C. pipiens and A. aegypti.

The relationship between initial population density of Globodera pallida and yield was examined for five genotypes of potato at four different experimental sites; two in Scotland and two in England. The experiments at each site were conducted in two stages. The experimental area was first manipulated using resistant and susceptible potato cultivars to give plots with different nematode population densities to act as the initial populations (Pi) for the second stage. The main experiment was conducted using a range of host genotypes that differed in their tolerance and susceptibility. The utility of an inverse linear model relating yield to initial population density was confirmed as were differences in cultivar tolerance. Site differences in yield loss were also observed. It was found that the model could be modified to include the partitioning of genotype and site effects. The implications of this in terms of developing further models to predict yield loss and nematode population dynamics are discussed.

Three double-marker systems were used in cannulated steers offered straw-based diets supplemented with barley, fishmeal or a combination of both, in an attempt to reduce the risks of misinterpreting biological results based on calculations of digesta flows using markers. Ruthenium phenanthroline (Ru-phe), ytterbium acetate (Yb-ac) and the indigestible acid detergent fibre fraction (IADF) were used as markers of the particulate phase of digesta in conjunction with Cr-EDTA as a liquid-phase marker to estimate duodenal and ileal digesta flows. The present paper presents the difficulties met in the application of marker techniques.

The particulate marker used in a double-marker system for calculation of nutrient flow modified the degree of statistical significance observed for dietary treatment effects on digesta flow parameters.

The data suggest that the type of diet used affects the behaviour of individual markers in different ways and could incur some bias in the results. It is hypothesized that the physical and chemical characteristics of digesta, in different dietary conditions, can be responsible for differential attachment of markers to particles of the solid phase. In such cases, the distinction made in the double-marker technique between particulate and liquid phases may not be sufficient to correct for lack of representativeness of digesta contents sampled from T-shaped cannulae.

For the diets studied it was concluded that the interpretation of digestion parameters should be based on both Ru-phe and Yb-ac results.

The initial work on the composition of the fluid obtained from the udders of virgin heifers has been extended and the character of the secretions of the mammary gland during first pregnancy has been investigated.

The results obtained point to the following conclusions:

During the first four or five months of pregnancy, the character and composition of the mammary secretion undergo no essential alterations. During this time, the mechanism of the gland which elaborates the normal constituents of milk is displaying feeble activity.

After about twenty weeks the secretion undergoes a progressive change in character and composition. The characteristic milk constituents begin to disappear, and ultimately the secretion, which is now honey-like in consistency, contains about 40 per cent, total solids, of which about 35 per cent, is globulin. During this phase, which apparently corresponds with a definite stage in the gland's development, the normal activity of the cells which results in the production of caseinogen, lactose, etc. seems to be in abeyance.

Removal of the honey-like secretion from the udder leads to a stimulation of normal mammary gland activity, in the same way as thorough and frequent milking stimulates activity during normal lactation. The viscous secretion gives way to a milk-like fluid. This is produced in relatively large amount and contains appreciable percentages of the normal colostrum corstituents.

If the honey-like secretion is not removed from the gland, the transition from this phase to the eclostral phase appears to be much more gradual and is probably taking place uniformly throughout the remainder of pregnancy.

Merino sheep were given abomasal infusions of either mixtures of amino acids or protein during periods of 8 or 12 days. Effects on wool growth were measured using autoradiography and a clipping procedure which allowed time for the emergence of the wool fibres. Estimates of volume growth rate, from the autoradiographic measurements, and of mass of wool grown, from clipping, were in good agreement.

An infusion of a standard mixture of 13 amino acids, which included ten essential amino acids in approximately the proportions in casein, consistently stimulated wool growth. The mean increases in volume and mass of wool grown, during 30 infusions, were 66 and 67% respectively. A mixture of ten essential amino acids alone appeared to be as effective as the standard mixture for stimulating wool growth, and there were no significant differences between the effects on wool growth of casein and the standard mixture of amino acids.

The omission of methionine from an infusion of the standard mixture of amino acids, or from a mixture of essential amino acids only, inhibited wool growth rate; both fibre diameter and length of wool grown per day were reduced to below the control values. In addition, the strength of the fibres was considerably reduced.

Infusions of zein and of an amino acid mixture simulating the essential amino acid composition of zein both inhibited wool growth rate, due to a reduction in fibre diameter. Similar effects on wool growth were observed when any one of three essential amino acids (lysine, isoleucine or leucine) was omitted from an infusion of the standard mixture of amino acids. The omission of five other essential amino acids (arginine, histidine, phenylalanine, threonine or valine) from the infusion, or variations in the proportions of leucine, lysine or methionine, had no appreciable effects on wool growth.

The paper seeks to compare the effectiveness of various functions as appropriate to express protein retention and lipid retention as a function of pig live weight. Linear descriptors were adequate for protein growth over the 20 to 120 kg live weight range, but not adequate when higher live weights were included. Linear descriptors were not adequate for lipid growth over any weight range. Quadratic and cubic polynomials, and the logistic function, were faulted on their failure to describe the known biology. Augmentation of the allometric function added nothing to the simple form. The Gompertz function was consistently effective for the description of daily protein retention rate. The Bridges and Richards functions are more flexible than the Gompertz with respect to the point of inflection, but when applied they resembled the Gompertz and therefore did not materially influence the descriptive outcome. The cubic polynomial, augmented allometric, Bridges and Richards functions, although favoured in other reports, were found to add nothing to the more simple functions. It is concluded that protein retention can be well expressed in relation to live weight by linear and Gompertz functions. Lipid retention could be well described by an allometric relationship with pig mass.

Peat soils change gradually as a result of drainage and cultivation, losing depth through shrinkage, oxidation and blowing, changing in character through increases in mineral matter and ‘drumminess’, and changing in productivity and manurial requirements. The sequence of changes has been studied in the field and a number of stages distinguished which may be recognized quite simply from the texture of the ploughed layer. Texture is a field assessment of the type and condition of the peat and the proportion and nature of the mineral matter associated with it. The Fenland soils have been classified as follows: (1) peat soils, (2) peaty mineral soils, (3) organic mineral soils and (4) mineral soils. Further subdivision into textural classes depends on recognizing the mineral combinations or fractions as in texturing mineral soils

Analysis of samples of the ploughed layer in the laboratory has shown that organic matter or loss on ignition figures, expressed as a percentage of the oven-dried soil, can be used for recognizing the kind of soil; use is made of this in considering those earlier experiments for which organic matter figures could be calculated.

Two groups of nine wethers, three of which were fitted with rumen cannulas, were used in a digestion trial at the INRA centre in Theix, France, in 1988. Group 1 received 65% maize silage and 35% concentrates; group 2 received 65% hay and 35% concentrates. Concentrates were based on either cereals rich in starch, or by-products rich in fibre and were given either alone or supplemented with lipids as calcium soaps. The fatty acid content of lipid-supplemented diets was c. 9·5 and 8·5% of dry matter, of which 85 and 89% was provided by calcium soaps, for maize silage and hay diets, respectively. For each group, the four diets were tested in four successive periods from January to June 1988.

Total digestibility of dry and organic matter, acid detergent fibre (ADF) and neutral detergent fibre (NDF) was measured in six wethers of each group by total faeces collection. On cannulated wethers, volatile fatty acid content and composition, pH and NH3-N in rumen liquor were determined four times a day; in sacco degradability of dry matter, ADF and NDF of the forage eaten by the wethers was estimated by the kinetics of incubation in the rumen.

In vivo and in sacco results showed that dry matter and organic matter digestibilities were not modified by the nature of concentrates. Cell wall digestibility was higher for fibre concentrates than for starchy concentrates, by 4·1 and 6·2 percentage units for NDF in maize silage and hay groups, respectively. Volatile fatty acids (VFA) and ammonia concentrations were higher and pH was lower with the maize silage than with the hay diet.

Lipid supply slightly increased cell wall digestibility in the group fed maize silage by 7·5 and 2·0 percentage units for starch and fibre concentrates, respectively. This surprising increase was related to an improvement in in sacco degradability. In all diets, lipid supply increased pH, but variations in VFA concentration and pattern were low. Interactions between the nature of concentrate and lipid supply were moderate, but were higher in the group fed maize silage than in the group fed hay, especially for total digestibility.

The relative importance of some sources of plant-to-plant root-weight variation in a field-grown carrot crop was investigated in two experiments sown in contrasting soil conditions but using the same seed lot and experimental design. In each experiment the following possible causes of root-weight variation were recorded on an individual plant basis: umbel order and size of seed, sowing depth, time taken for seedling emergence, length of cotyledons at emergence, distance to nearest neighbours within a row and row position within the seed bed. The relative importance of these differing sources of root-weight variation was assessed by determining how much of the variation in In (root dry weight) they accounted for at two harvests in each experiment.

In the first experiment, carrot seeds were sown in warm, moist conditions in May. These conditions gave a near-ideal emergence, 82% of seedlings emerging over only 5 days. Thirteen days after sowing, the most important source of root-weight variation, out of the variables measured, was position of the row within the bed. However, approximately two-thirds of the root-weight variation could not be attributed to any of the measured variables.

In the second experiment carrot seeds were sown in hot, drying conditions in July. The subsequent spread of seedling emergence extended over approximately 40 days. The total sum of squares of In (root dry weight) in this experiment was six times greater than the sum of squares in the first experiment. At 105 days, 61% of the rootweight variation in Expt 2 was accounted for by differences in time of seedling emergence. Differences in cotyledon length, measured when fully expanded, accounted for 24% and sowing depth differences accounted for 19% of the root-weight variation. Differences in umbel order, seed size and distance to the nearest neighbouring plants within a row never accounted for more than 8% of the sum of squares of In (root dry weight) in either experiment. Thus these variables were not important sources of rootweight variation. In contrast variables which affected the time of seedling emergence, such as sowing depth and physical conditions around the seed, were major sources of root-weight variation.

Safety evaluation of a genetically modified (GM) crop is accomplished by establishing its substantial equivalence to non-GM reference crops with a history of safe use. Testing hypotheses of equivalence rather than difference is the appropriate statistical approach. A necessary first step in this regard is to specify a reasonable equivalence criterion that includes a measure for discrepancy between the GM and reference crops as well as a regulatory threshold. The present work explored several equivalence criteria and discussed their pros and cons. Each criterion addresses one of three ordered classes of equivalence: super, conditional and marginal equivalence. Their implications were investigated over an array of parameter values estimated from a real-world dataset. Marginal equivalence was identified as adhering most closely to the concept of substantial equivalence. Because conditional equivalence logically implies marginal equivalence and is practically quantifiable from current field designs, the present work recommends conditional equivalence criteria while encouraging producers to improve their design to enable testing marginal equivalence in the future. Contrary to concerns of the ag-biotech industry, empirical evidence from recent publications indicates that a linear mixed model currently implemented by the European Food Safety Authority is adequate for assessing equivalence despite its lack of genotype-by-environment interaction terms.

There was a significant positive correlation between the buffer capacity and motility of sperm and a highly significant negative correlation between the buffer capacity and the pH. There was no correlation between the buffer capacity and the volume of the ejaculate or the number of sperm per c.mm. Specimens which maintained a good motility on storage had a better buffer capacity than those which did not.

The specific gravity of 125 semen specimens was 1·0337. There was a highly significant positive correlation between the specific gravity and the number of sperm per c.mm. and the percentage motility and a highly significant negative correlation between the specific gravity and the pH. There was no relationship between the specific gravity of semen and the volume of the ejaculate, or the pH change. In general, the higher the specific gravity of the seminal fluid, the better was motility maintained on storage. The higher the specific gravity, the lower was the pH and the better the buffer capacity of seminal fluid.

At the research farm of the Swedish University of Agricultural Sciences, Uppsala, above- and belowground production and changes with time in root length, mean root radius and root surface area of spring oilseed rape were studied during the growing seasons 1987 and 1990. In both years, the highest root growth rate was recorded during the stem elongation phase, and the highest shoot growth rate during flowering. The root: shoot ratio decreased throughout the whole period of root sampling, from 0·64 to 0·16, during the cool and wet first year. In the warmer and drier second year, the ratio increased to a maximum of 0·72 when flowering started, and thereafter decreased. More than 80% of the root dry matter was found in the topsoil. Roots were longer and thinner in the dry and warm 1990 than in the wet and cool 1987. Maximum root length was c. 4·9 km/m2 in 1990, and mean root radius varied between 01 and 0·7 mm. Increases in root surface area during periods of root growth were due to increased root length rather than to increased mean root radius.

The hardseededness and pattern of hard seed breakdown of 35 accessions of Medicago polymorpha, collected in Sardinia, were compared with Circle Valley, an Australian cultivar, near Sassari, northern Sardinia. The seed was produced in three consecutive years (1991–93), and the hardseededness measured at intervals in trie field during the summer following its production. The results were related to previously measured pod characteristics that separated the varieties polymorpha and vulgaris

The variety polymorpha, which had been collected mainly from coastal regions, was harder seeded than the variety vulgaris, which had been collected mainly from cooler, inland regions. Wherever the origin, hardseededness levels remained > 85% even in the year when seed softening was greatest (1992). In other years means of 94% (1991) and 97% (1993) of seed remained hard in late September. In 1992 and 1993 most hard seed breakdown occurred between early August and early September, while in 1991 hard seed breakdown occurred in September itself. There was little difference between varieties in the pattern of hard seed breakdown. Seedling emergence was related to the amount of seed that became soft during the previous summer. The results also confirmed that hard seeds soften sequentially from the proximal end of the pod (nearest the calyx) to the distal end.

The Australian cultivar was one of the hardest seeded accessions, showing very little hard seed breakdown during the experimental period. This result highlights the importance of selecting pasture legumes under the conditions in which they are to be used: in Australia, cultivar Circle Valley usually loses some 20% of its hardseededness in the first year, comparable with the softest Sardinian accessions in Sardinia. Even among the Sardinian accessions, most were probably too hard seeded in 1993. Nevertheless it is likely that material in this experiment will be released as commercial cultivars for Sardinia and other similar environments in the Mediterranean basin.

1. An outline of recent investigations on the cause of marsh spot in pea seeds indicates that the disease is caused by a deficiency of available manganese in the soil.

2. The symptoms of the disease are very briefly described.

3. The methods of analysis used by the writers for determining the amounts of manganese in whole peas and in different parts of peas are described.

4. As no loss of manganese was found to result from soaking the peas in water for 24 hr., soaked peas were used for dissection in preference to dry peas.

5. Referring to the diseased peas, the highest level of manganese was found in the peripheral tissues of the cotyledons, followed by the germ and seed-coat. Only slight traces of manganese were found in the healthy and in the necrotic tissue from the centre of the cotyledons. All parts of the healthy seed were richer in manganese than the corresponding parts of the diseased sample. In addition, the order of distribution was different; notably in that the tissue from the centre of the cotyledon contained a higher level than the seed-coat, and the germ a higher level than the outer layers of the cotyledon.

6. A given weight of small peas proved to contain less manganese than the same weight of large peas selected from the diseased sample. The reverse was found to be true for similar size groups selected from the healthy sample. These differences in the manganese content of size groups suggest that peas of uniform size should be selected for analysis when the relative manganese levels of different samples are to be tested.

7. It is suggested that migration of cell contents from the necrotic tissue of diseased peas may partly account for the differences in manganese content of healthy and diseased peas.

The concentration of glutamate and of glutamine was measured in whole blood obtained from a maternal artery, a uterine vein, a foetal artery and an umbilical vein of chronically catherized ewes and foetuses from 100 to 140 days after conception. The activities of glutamate dehydrogenase, phosphate-dependent glutaminase, phosphateindependent glutaminase, glutamine synthetase, γ-glutamyl transferase and glutamine-oxo-acid aminotransferase were measured in placentomes obtained from ewes during a similar period of gestation.

The concentrations of glutamate in blood from maternal vessels remained constant, whereas there was a significant decline (P < 0·001) in the concentration of glutamate in foetal blood. Glutamine concentrations declined significantly (P < 0·05) in maternal blood and in foetal arterial blood (P < 0·001), whereas the concentration of glutamine in umbilical venous blood remained constant.

Mean arterio-venous differences for glutamate indicated that there was no net uptake from or release into maternal blood by the uterus. However, there was a significant (P < 0·02) uptake of glutamate by the placenta from the foetal circulation. Glutamine release from the placenta into the foetal circulation increased as the foetus matured.

Significant activities of glutamate dehydrogenase, γ-glutamyl transferase, glutamine synthetase and phosphate-dependent glutaminase were found in the placenta but there was no significant relationship between the activities of these enzymes and the gestational age of the foetus. The enzyme profile indicated that the placenta has a substantial potential for net glutamine synthesis.

It is concluded that, for a 140-day foetus, the release of glutamine from the placenta accounts for more than half of its nitrogen requirement. Direct placental transfer of glutamine from maternal blood accounts for only one-third of the glutamine released by the placenta into the foetal circulation of a 140-day foetus. Therefore, the remainder of the glutamine is synthesized in the placenta from glutamate. Only one-third of the glutamate required for this placental glutamine synthesis is from the glutamate released by the foetus. The remainder must be derived either from 2-oxoglutarate, as the result of aminotransferase or glutamate dehydrogenase activities, or from glutathione by the action of γ-glutamyl transferase.

Four perennial ryegrass varieties, S.24, S.321, S.23 and Vertas Hardgrazing, were grown in micro plots containing 30 plants each at a spacing of 5 cm apart. Monocultures and four of the possible six binary mixtures between the varieties were grown in a randomized block design with four replicates. The data recorded comprised dry-matter yields and tiller numbers, and the components of the mixtures were harvested separately. The data were based on both whole plots, to give total production of mixtures and monocultures, and half plots, to give the contribution of mixture components to total mixture production. Four harvests were obtained during 1968, and the data from these were analysed separately and summed to provide the annual total data.

Mixtures and monocultures did not differ significantly for annual total dry-matter yield or for dry-matter yield at the separate harvests. However, differences in drymatter yields between monocultures and between mixtures were detected at some harvests. Within mixtures, significant differences between some of the respective components were found at certain harvests, but these were not obtained for the annual total data.

The stability of monocultures and mixtures was investigated by a regression technique and considered in the context of the environments used. This technique revealed that there was a specific mixture effect whereby the mixture (S.24 + S.23) differed significantly from all other treatments.

Significant differences between monocultures for fertile and sterile tiller numbers were detected. Between mixtures there were differences for fertile tiller numbers, and within mixtures certain of the components differed for fertile, sterile and total tiller numbers.

The relationship between dry-matter yield and tiller numbers was investigated by regression analyses over all treatments and for monocultures and mixtures separately. Significant differences between the slopes of the regression lines and zero and between the slopes of the regression lines themselves were obtained.

Possible reasons for the differences recorded both between and within monocultures and mixtures are considered and related to the competitive relations existing in the trial.