Recently developed regressional techniques suggest that some of the problems posed by genotype-environment interactions may be solved. These techniques have been applied, therefore, to data from an extensive trial involving perennial ryegrass (Lolium perenne) varieties grown under various treatments at several locations throughout Great Britain. Analysis reveals that the relative performance of the varieties depends mainly upon whether they are grown as spaced plants or swards. Even though the genotype-environment interactions are large, much of their effects could be reduced to an essentially linear scale. These results are considered in the light of the known characteristics of the varieties concerned, whilst the wider implications are also discussed.

The live-weight gains of male and female lambs were measured during growth from 16 to 40 kg live weight when five cereal-based diets varying in dietary crude protein concentration (from 10 to 20%) were given at three levels in a 5×3×2 factorial design.

Males grew faster than females by about 15% (P < 0·001); this difference increased with age and with the amount of feed given (P < 0·05). Growth rate responded linearly to increase in feeding level (P < 0·001) and curvilinearly to increase in dietary protein concentration (P < 0·01). A significant interaction (P < 0·01) occurred whereby growth increased with higher protein concentrations as feeding level increased.

At the highest feeding level (near ad libitum) the results suggest that the optimum dietary crude protein concentration for growth was about 17·5, 15·0, 12·5 and 12·5% at body weights of 20, 25, 30 and 35 kg respectively. The overall optimum dietary crude protein concentration for growth between 16 and 40 kg body weight was about 17·0, 15·0 and 11·0% when the mean digestible energy intake was 3·0, 2·6 and 2·1 Mcal/day.

The results agree reasonably well with estimates of protein requirements for lambs given by the Agricultural Research Council (1965).

1. In an experiment with ninety-nine lambs the effects on the body composition of male and female lambs were examined when five diets containing different concentrations of crude protein (in the range 10–20%) were given at three levels of feeding and lambs were slaughtered at two live weights (27·5 and 40 kg).

2. With lambs slaughtered at 27·5 kg there were significant increases in the rate of both nitrogen and fat retention with increases in levels of feeding. There were also linear increases in the rate of protein deposition and decreases in fat deposition with increases in the concentration of crude protein. This effect was particularly marked at the high level of feeding.

3. With lambs slaughtered at 40 kg live weight there were also linear increases in fat and in nitrogen deposition with increasing feeding level but the effect of increasing the protein concentration on increases in nitrogen retention departed from linearity.

4. While at 27·5 kg there were no significant effects of feeding level on nitrogen and ether-extract content of the bodies at slaughter, with animals slaughtered at 40 kg there was a significant linear decrease in ether-extract content with increasing feeding level and a corresponding linear increase in nitrogen content with increased level of feeding.

5. Male lambs deposited more nitrogen and less fat than females. This was true of both rate of deposition and of carcass composition at 40 kg live weight.

Modifications to the New Zealand meter (Campbell, Phillips & O'Reilly, 1962) are briefly described and the results obtained with the modified meters presented.

Water was shown to have the dominant influence on meter reading—dessication of the plant material reduced readings, application of water increased readings.

The relation between yield of water (W) and meter reading (MR) depended on the circuitry of the meter used and was similar to that between capacitance change induced by a variable capacitor and meter reading. With the new meters this relation was strongly linear. For individual species the simple correlation of W on MR was generally superior to that for weight of dry matter. Regression coefficients for weight of water on meter reading differed between species and/or days but were correlated with the relative water content of the species (RW). Inclusion of RW in addition to that of W increased the variation accounted for (V%) when all species were included from 76·0 to 86·9%. However, the multiple regression did not eliminate bias. For individual species, there was no benefit in including the RW term.

The inclusion of height of herbage (H) in a multiple regression equation with weight of water greatly improved the V% when compared with the regression of meter reading on weight of water alone for oats dressed with three levels of nitrogen. The V% increased from 71·3% for W alone to 90·0% when W, H and H2 were included in the regression. Further testing using height of herbage over a wide range of species will be required to confirm this improvement for use in an overall regression.

Emphasis is placed on the very strong linear relation between yield (of water or dry matter) and meter reading within any one pasture at any one time. Use is made of this relation in the presentation of a new method, which overcomes the bias associated with prediction equations, for estimating pasture yields. This entails establishing the mean reading for any pasture and cutting for yield estimation three samples having such a mean. The limitations of the method for use with mixtures is discussed.

Nine experiments in the East of Scotland were completed in the years 1966, 1967 and 1968 to test the response of potatoes grown for ware to five levels of a 15½–15½–21 fertilizer applied at rates of 0, 4, 8, 12, 16 and 20 cwt per acre.

1. Mean yield of ware tubers increased up to the highest fertilizer level.

2. Mean yield of small tubers increased only up to the 8 cwt per acre level and then declined.

3. Dry matter content of ware tubers decreased with increasing levels of fertilizer.

4. Increasing fertilizer levels lowered soil pH. At one site this was associated with a significant reduction of common scab on the tubers.

Differences between Sahiwals and Jerseys and between summer and winter in the quantities of lipids on the skin surface and on hair were studied.

Sahiwals had significantly greater quantities of lipids on both the skin surface and hair in summer and in winter. In both breeds, there were significantly greater amounts of lipids on the skin surface and on hair in winter. In both breeds a significantly greater quantity of lipids was present on hair than on skin in winter but not in summer.

Jerseys had greater weight of hair/unit area of skin than Sahiwals in both summer and winter. There was no significant correlation between the amount of lipids on hair and the weight of hair.

The digestibility of some fats (beef tallow, HEF, herring oil, soyabean oil and maize oil) were studied in balance trials with sheep which had been maintained on the diets for a 35 day period of acclimatization to establish ‘steady state’ conditions. Corrected digestibility coefficients for the individual fats were calculated from a knowledge of the intake and output of lipid on a low-fat basal ration. The difference between the intake of fat and the corrected lipid output (total lipid output minus lipid output on basal ration) expressed as a percentage gave the corrected digestibility coefficients.

It was found that the corrected digestibility coefficients for beef tallow was 85%; HEF, 74%; herring oil, 84%; soyabean oil, 83%; and maize oil, 70%. The values for beef tallow and HEF were comparable to similar estimates in pigs. The digestibility of the highly unsaturated soyabean and maize oil were markedly lower in comparison with non-ruminant species. Extensive hydrogenation of dietary unsaturated fatty acids was confirmed. For soyabean and maize oil the corrected output of stearic acid exceeded the dietary intake.

The effect of fatty acid chain length and unsaturation on digestibility in sheep were examined using partially purified samples of lauric, myristic, palmitic, stearic, oleic and linoleic acids. The digestibility of the fatty acids was relatively constant with only a very slight decrease on increasing chain length. There was an extensive hydrogenation of the unsaturated fatty acids.

The corrected digestibility coefficients for lauric acid was 91%, myristic 86%, palmitic 87% and stearic acid 81–83% whereas the corrected digestibility coefficients for oleic and linoleic acids were calculated at 87 and 93% respectively. The digestibility coefficients for the saturated fatty acids are higher than similar estimates that have been reported for non-ruminants. It is suggested that the ruminant is better able to utilize saturated fatty acids than the non-ruminant.

Four soil moisture and two nitrogen levels were combined in eight treatments to study moisture and nitrogen effects on Dular (June–September crop) and Patnai-23 (June–November crop) varieties of rice. Both the varieties when grown under natural precipitation yielded significantly less than at other moisture levels. Whereas there were no significant differences amongst the three treatments of submergence (15–10 cm), cyclic submergence (15–0 cm) and cyclic wetting-drying (saturation to field capacity) for Dular rice, the former two treatments proved better than the latter for Patnai-23 rice. Dular responded to the treatments of cyclic wetting-drying in a way similar to that of submergence because the crop completed its life cycle during the period of low evaporative demand. Among the levels of nitrogen, 40 kg N/ha proved optimum for Dular rice and 6O kg N/ha for Patnai-23 rice. The increase in yield of grain was associated with increase in the number of panicles, number of spikelets/panicle and percentage of ripened spikelets. The concentration of N, P, Fe and Mn in the plant was maximum under submergence followed by cyclic submergence, cyclic wetting-drying and natural precipitation. Higher levels of nitrogen increased only N and P contents in the plant.

Results are given for a second period of 9 years, 6 on the pattern of 1951–9, with the last 3 years used for ad hoc management experiments. The meteorological specification of a year of ‘irrigation need’ was satisfied in 4 years: there were good responses in 5 years. During the first 6 years, out of twenty-four sets of crop yields (from eight different crops), very good responses were obtained on nine, good responses on eight, and zero or very slightly negative responses on the other seven.

A general theory of the inter-relationships of growth and water, and of growth and radiation, leads to an expression

k = 39ε t ha−1 cm−1,

as the maximum possible response to irrigation in terms of total dry matter produced, where ɛ is the fraction of solar radiation fixed by the crop (ɛ has a range 50–100 × 10−4 for good to very good farming) when irrigation is needed. The ‘need’ is defined through

Y = k(ET + De − Dm),

where Y is the yield, ET is the total potential evaporation, De is a limiting deficit up to which there is no check to growth, and beyond there is no growth, and Dm is the maximum deficit experienced by the crop at the time of measuring Y. The agronomic object of the experiments is to determine De, i.e. what reserve of soil water can be used by the crop before irrigation is necessary.

Experiments on a sandy loam, using adequate fertilizer, gave:

(a) Crop yield (as total dry matter) when soil water content was kept near field capacity.

(b) A measure of the departure from field capacity (limiting deficit, Dι) that produced no detectable change in yield.

(c) Responses to irrigation: (i) conventional, as δY/δI; (ii) a theoretical maximum, k, after allowing for Dι in the water balance.

(d) Other information, botanical and technical.

Grass (S 22 Italian ryegrass)

Two year crop, cut fourteen times, October 1959 to October 1961.

(a) Best yields: 1960, 15 t ha−1 (ɛ ≃ 100 × 10−4); 1961, 12 t ha−1 (ɛ ≃ 70 × 10−4).

(b) Dι ≃ 5 cm.

(c) (i) δY/δI: 1960, 0·36; 1961, 0·43 t ha−1 cm−1; (iii) k: 1960, 0·40; 1961, 0·27 t ha−1 cm−1.

(d) In 1960 irrigation was needed in spring only: the benefit was maintained on plots with large N and K dressings, but almost disappeared on those with small dressings (Y = 11 t ha−1). In 1961 the unirrigated crop could not exploit all the limited summer rain: hence δY/δI > k.

One year crop, cut five times, March to October 1965.

(a) Yield, 12 t ha−1 (ɛ ≃ 90 × 10−4, from emergence).

(b) and (c) (i) The meteorological need was slight, Dι (above) was not exceeded, and there was no response; (c) (ii) k = 0·46 t ha−1 cm−1.

Lucerne (Du Puits)

Three year crop, cut eleven times, April 1962 to November 1964.

(a) Best yields: 1962, 7; 1963, 8; 1964, 10 t ha−1.

(b) Dι = 11 cm for established crop.

(c) (i) No response other than in first dry spring of establishment. At first cut δY/δI = 0·18 t ha−1 cm−1, (ii) k = 0·26 t ha−1 cm−1.

(d) Extra K seemed to be helpful in sustaining the early benefit throughout the 3 years.

Clover (Crimson)

One year crop, drilled April and cut (once only) July 1963.

(a) Best yield: 3 t ha−1.

(b) Dι, not detectably different from zero (evidence inadequate).

(c) (i) δY/δI = 0·16 t ha−1 cm−1; (ii) k ≃ δY/δI (?).

(d) First cut completely defoliated irrigated crop, there was no recovery, and the experiment was abandoned.

Clover (Dorset Marl)

One year crops undersown in previous barley crops, 1963 for 1964 (three cuts), 1964 for 1965 (three cuts).

(a) Best yields: 1964, 9 t ha−1 (ɛ ≃ 60 × 10−4); 1965, 8 t ha−1 (ɛ ≃ 60 × 10−4).

(b) Dι ≃ 2·5 cm.

(c) (i) δY/δI: 1964, 0·24 t ha−1 cm−1. 1965, no response; (ii) k: 0·22–0·34, depending on method of derivation.

(d) The response to irrigation was slightly greater on plots not previously irrigated in the barley year. In (c) the uncertainties arise out of excessive rain after the first irrigation in 1965; there was estimated drainage from the irrigated plots but not from the control plots. In 1964, the control plots failed to exploit all the summer rain.

General

There seems to be a need to know two limiting deficits, one, the Dt of these papers, that is a measure of what can be tolerated without any check to growth, and a second that is a kind of maintenance deficit, necessary to keep the crop non-senescent so that it can use all the rain it gets.

Irrigation greatly helped the establishment of all three leys, and both grass and clover responded well to later irrigation. Lucerne did not. In English units, the limiting deficits, Dt, for well-fertilized established crops were, approximately, 4 in for lucerne, 2 in for grass, and 1 in for clover.

Experiments on a sandy loam, using adequate fertilizer gave:

(a) Crop yield when soil water content was kept near field capacity.

(b) A measure of the departure from field capacity (limiting deficit, De) that produced no detectable change in yield.

(c) Responses to irrigation: (i) conventional, as δγ/δΙ (ii) a theoretical maximum, k, after allowing for De in the water balance.

(d) Other information, botanical and technical.

Sugar beet

(a) Best yields of sugar: 1963, 9; 1963, 10; 1964, 8 t ha-1.

(b) Limiting deficit can increase to ca. 10 cm by end of September.

(c) (i) Best δγ/δΙ. 1963, 0·18 t ha-1 cm-1 (some need); 1964, 0·62 t ha-1 cm-1 (great need); 1965, slightly negative (no need); (ii) inferred k = 0·4 t ha-1 cm-1 for sugar, and ca. 1·0 t ha-1 cm-1 for total dry matter. For the period sowing to harvest the values of e are 0·7 and 1·8 x 10-2 respectively.

(d) Excess water (as in 1964) seems to depress the sugar yield slightly.

Potatoes, main crop (part of an experiment on control of nematodes, under continuous cropping)

(a) Best fresh weight yields of tubers: 1966, 46 (without irrigation); 1967, 38 (with irrigation); 1968, 33 (with irrigation), t ha-1.

(b) Evidence inadequate.

(c) Responses were small, or negative; (i) best response, in 1967, 0·7 t ha-1 cm-1; (ii) evidence inadequate.

(d) Decreased yield in 1966 may be caused by early summer leaching: the effect was about equal to removing half of the nitrogen fertilizer applied.

Potatoes, early (Arran Pilot)

(a) Best fresh weight yields of tubers: 1960, 32; 1961, 37; 1962, 16 t ha-1, all at larger of two nitrogen dressings, and after normal cultivation.

(b) De ≃ 2·5 cm.

(c) (i) Best δγ/δΙ. 1960, 1·7; 1961, 2·0; 1962, 1·1 t ha-1 cm-1, (ii) k = 1·8 t ha-1 cm-1 for nitrogen at 0·6 cwt acre-1; k = 2·11 ha-1 cm-1 for nitrogen at 1·2 cwt acre-1. Inferred values of ɛ are 1·15 and 1·45 x 10-2.

(d) Weed control by chemical means was not successful: all yields and responses were decreased by about one-third.

Barley, spring (Proctor, Maris Badger)

After early potatoes, some plots has a crop of trefoil, ploughed in (crops of 1961–3). In 1968, barley was test crop in an experiment on fumigants. Cropping in 1960 and 1964 was normal.

(a) Best yields of grain, 3·4–4·8 t ha-1 (without trefoil); 4·1–5·4 t ha-1 (after trefoil).

(b) De ≃ 4 cm before ear emergence, and unimportant thereafter.

(c) (i) Best δγ/δΙ, in 1962, ca. 0·20 (no trefoil), and 0·30 (after trefoil) t ha-1 cm-1, both at upper of two nitrogen dressings; (ii) k (for variety Proctor only, and no trefoil) ≃ 0·20 t ha-1 cm-1, and may be independent of nitrogen treatment.

(d) Previous management of early potatoes did not affect barley yields. For 1961–3—average yield of all treatments = 3·53 t ha-1—the average responses were: to trefoil, 0·46 t ha-1; to nitrogen, 0·76 t ha-1; to water 0·95 t ha-1. Trefoil halved the response to nitrogen, and may have increased the response to water.

Within limits, water and nitrogen seem to be interchangeable as management factors.

Wheat, spring (1965, Opal, in a normal experiment. 1966, 1967, Kloka in an experiment with a dwarfing compound, CCC. Four levels of N fertilizer)

(a) Best yields of grain, 1963, 4·5; 1966, 5·8; 1967, 5·9 t ha-1.

(b) De ≃ 4 cm before ear emergence, and ca. ET/4: thereafter.

The evidence is flimsy.

(c) (i) Best δγ/δΙ. 1965, 0·11; 1956, 0·16; 1967, 0·12 t ha-1 cm-1; (ii) evidence is inadequate to derive values of k, but it seems to be very small with N1 (0·4 cwt acre-1 N), and perhaps of order 0·2 t ha-1 cm-1 at all the bigger dressings (2–4 x N1).

(d) On average, CCC had no effect on grain yields, but it decreased straw yields. In 1967 many sub-plots gave very large yields: the average of nine was 12 t ha-1 as total dry matter; three were not irrigated, and had no CCC; three had CCC and were irrigated. None was at the smallest N treatment. Spring wheat responds to water like any other crop when given enough N (ca. 0·8 cwt acre-1 N, or a little more). The need continues after ear emergence.

Beans, winter 1960, spring 1961, 1962, 1968

(a) Best yields of grain, 1960, 3·7; 1961, 3·1; 1962, 3·6; 1968, 3·0 t ha-1.

(b) De ≃ 4 cm.

(c) (i) Best δγ/δΙ. 1960. 0·18 (late watering); 1961, 0·15 (early watering); 1962, 0·22 (late watering); 1968, 0·02 (early watering and then rain in excess); (ii) k = 0·17 t ha-1 cm-1.

(d) There is no evidence to show that irrigation is more important either before or after flowering (early and late, in (c)).

General

Exception for the main-crop potatoes, all crops responded to irrigation as expected, with gains of 50–100% in the driest summers. In English units, the values of the limiting deficits were, approximately, up to 4 in by the end of September for sugar beet, 1 in for early potatoes, 1½ in for barley before ear emergence, 1½ in for wheat before ear emergence and up to about 3 in 5 weeks later, and 1½ in for beans.

Seed of Lolium perenne L. was grown from material collected from three S.23 leys and one S.24 ley that had been subjected to high rates of stocking and nitrogen application. Comparisons were made between plants grown from seed derived from the leys and from breeders' seed of S.23 and S.24.

Seedling characters indicated that the populations derived from leys differed from each other and from the plants grown from breeders' seed. In S.23 populations the main change occurring in intensively grazed leys was towards higher tillering prostrate seedlings. This did not occur in the S.24, but here the survivor population had a slightly shorter fifth leaf and a lower number of leaves on the main tiller of the seedlings.

As spaced plants, the population derived from the S.23 ley, with the highest percentage ryegrass and the reputation for the highest yields on the farm, gave a higher yield, recorded in three cuts in the first year and in four cuts in the second year, than any other populations. Small differences between S.23 populations in aftermath heading, reaction to rust, and mean date of panicle emergence were recorded. Variance of spaced plant yield decreased in ryegrass derived from older S.23 leys. The S.24 ‘survivor’ population was much later in mean date of panicle emergence than S.24 grown from breeders' seed.

A Nigerian, long-season, sorghum produced 15 tons of dry matter/acre in about 155 days, but the grain yield was small (2000 lb/acre or 2242 kg/ha) because only part of the dry matter formed after heading went into the grain. In contrast, an American, short-season hybrid which formed only from one-half to one-third as much dry weight as Farafara, gave over 4000 lb of grain/acre (4484 kg/ha) in 100 days. A Nigerian sorghum of intermediate height and a long-season hybrid, produced by crosses between Nigerian varieties and an American, short-season sorghum, gave larger yields than the tall variety because they formed more heads/unit area and more grains/head.

Crop growth rates of the three long-season sorghums sown in June increased to a peak in July, but declined sharply to very small values in September, and then increased again very rapidly at the end of September as heads emerged. The most important factors causing the decline were a decrease in radiation and a loss in dry weight from decay and detachment of dead leaves. An increase in radiation at the end of September and developmental changes associated with head emergence contributed to the increase in growth rates just before heading.

A large proportion of the dry weight formed after heading in Farafara accumulated in the stem but a smaller proportion in the other long-season variety and the longseason hybrid. There was an inverse relationship between the increase in stem weight and the yield of grain. For large grain yields, sorghums are needed that will produce heads with sufficient grains to accommodate all the assimilate that the leaves produce at the end of the rains when crop growth rates are increasing.

Changes in the leaf area and light profiles of a tall and a short sorghum are described and the development of leaf area in these and in two other sorghums of intermediate height is examined in relation to the rate of appearance, the size and the length of life of the leaves. The extent to which these changes are related to the seasonal fluctuations in crop growth rate is discussed.

Pregnant ewes of the Scottish Blackface and Romney breeds were group fed to achieve different patterns of live-weight change between mating and parturition. One group had a net maternal live-weight increase of about 8%, the second a net loss of about 5% and the third a loss of about 16%. Wool growth and sulphur content of the wool were measured at intervals. During the period 6–8 weeks prepartum wool growth rate in the Blackface breed was low and there were no significant differences amongst treatments. Differences in sulphur content were highly significant (ranging between 3·97 and 3·46% for fine and coarse fractions respectively in the well-nourished sheep down to 3·68 and 3·20% in the undernourished group). In the Romney breed the situation was reversed. In the prepartum period there were no significant differences amongst groups in sulphur content (mean value 3·68%), but differences in wool growth rate were highly significant (0·021, 0·012 and 0·009 g/100 cm2/day).

Fifty-three experiments made between 1959 and 1968 tested the response to magnesium sulphate by sugar beet on fields where magnesium deficiency symptoms were expected. Soil samples, taken before applying fertilizers, were analysed for exchangeable magnesium by four methods. Sodium, potassium and calcium in the soil extracts were also measured to determine whether they influenced response to magnesium.

Results of different methods of analysing soil for magnesium were related to each other and to the percentage yield-response to magnesium fertilizer. The concentration of other soil cations did not affect response to magnesium fertilizer, but giving other cations, especially sodium, as fertilizer decreased the concentration of magnesium in the crop. Nevertheless, even on fields deficient in magnesium, the largest yield was from plots given sodium and posassium fertilizer together with a dressing of magnesium.

Sugar beet grown on soils containing less than 20 p.p.m. Mg extracted with ammonium nitrate usually gave a profitable response to magnesium fertilizer. When soil magnesium was 20–35 p.p.m., yield of sugar beet on some fields was increased slightly. Plants in some experiments had poorly developed root systems and response to magnesium was then always larger than expected from soil analysis.

Selfed and diallel progenies of selected clones of Lolium perenne were subjected to four nitrogen rates—0, 225, 450 and 675 kg/ha of N per annum. The herbage was analysed for total nitrogen (N), digestible dry matter (DMD), digestible organic matter (D) and acid-pepsin solubility was determined on total (PS) and organic matter (POMS) in 1967 and 1968.

Data for the percentage composition and the yield of quality components are presented with the variances of general (g.c.a.) and specific (s.c.a.) combining abilities for these characters. The g.c.a. and s.c.a. interaction with nitrogen is partitioned.

For percentage quality components and yields there were significant effects of years, nitrogen rates, progenies and their interactions. The variance for g.c.a. for all percentage quality components was not significant in 1968 but in 1967 all were significant except for total N. For yields, only the g.c.a.'s for PS and POMS were significant in 1967: no g.c.a.'s were significant in 1968.

There were considerable differences between years for heritability values. For digestibility the predicted performance of a theoretical F2 synthetic was little better than the mean of the population from which it was derived.

The implication of these studies in the formulation of a grass breeding programme is outlined.

Five selected clones of Lolium perenne were selfed and the ten possible diallel crosses, without reciprocals, made. The progenies were grown in a split-plot design in which whole plots comprised the four nitrogen rates (0, 225, 450 and 675 kg/ha of N) and sub-plots the fifteen progenies, square planted at 0·25 m.

The percentage of the applied nitrogen recovered, and the utilization of the nitrogen by the plant, were measured. Data were collected from four harvests in each of two consecutive years. The diallel was analysed by Griffing's (1956) Method II, Model II.

There were significant effects for nitrogen rates, progenies and their interaction for both ‘recovery’ and ‘utilization’, but years were significantly different for ‘recovery’ only. The effect of applied nitrogen was mainly linear and it was the linear effect of applied nitrogen which interacted most with the progeny effects.

Significant g.c.a. and s.c.a. effects were found but they varied between nitrogen rates and years. For ‘utilization’ the interaction nitrogen rates x g.c.a. was significant on two year means but there was no significance for the nitrogen interaction with s.c.a. The g.c.a. values obtained for both parameters did reveal some trends between the five parents.

The presence of significant additive and non-additive genetic variation is discussed in relation to a plant breeding programme.

Measurement of the variation in reproductive characters between genotypes of Phleum pratense showed that inflorescence length had a significant variance and was correlated with seed yield/plant. Genotypes with the longer inflorescences showed a quicker response to a long photoperiod and had a longer period of inflorescence development, and there was an indication that selection for increased inflorescence length might affect the production of leaves. A genetic analysis of a 6 x 6 half diallel cross showed that additive and non-additive effects were evident in the control of inflorescence length under natural conditions, but the non-additive effects were almost entirely suppressed under glasshouse conditions.