Two experiments were carried out to determine endogenous losses and the response of urinary purine derivatives to increased duodenal inputs of purine bases. Four ewes each fitted with a re-entrant cannula at the proximal duodenum, and conventionally fed, were subjected to full replacement of duodenal digesta followed by the administration of a solution either free of purines (Expt 1) or enriched with increasing amounts of purines, to supply 0·48–21·27 mmol/animal per day (Expt 2). Basal daily urinary excretions of allantoin, uric acid, hypoxanthine and xanthine were 11·5 ± 0·94, 9·9 ± 0·67, 6·9 ± 0·46 and 1·2 ±0·16 mg/kg W0·75. Allantoin was the only purine derivative which increased in response to incremental inputs of duodenal purines. The relationship between allantoin excretion and infused purines showed a urinary recovery of 0·8 for purines infused at > 220 μmol/kg W0·76. Lower rates of infusion did not alter allantoin excretion. The results show urinary allantoin to be a useful index to estimate duodenal input of purines when animals are fed close to or above their energy maintenance requirements.

For over 40 years a collaborative network of publicly funded international wheat scientists has made a significant contribution to food security in the developing world. Thousands of modern wheat varieties (MVs) have been released for use in both favourable and marginal environments on well over 50 million hectares. The yield increases associated with genetic improvement in yield potential and adaptation to biotic and abiotic stresses are well documented. Millions of small-scale farmers in the developing world have benefited. While this so-called ‘Green Revolution’ displaced landraces in favour of more productive MVs, these and other genetic resources, held in trust by international organizations, have been utilized to improve the inherent genetic diversity of modern varieties. Furthermore, the result of increased yields reduced the need to bring natural ecosystems under cultivation, by as much as a billion hectares.

Although international wheat breeding has its origins in the 1940s, recognition of a common scientific basis of agricultural problems worldwide was highlighted by the creation of International Agricultural Research Centres (IARCs) which included the International Maize and Wheat Improvement Centre (CIMMYT) established in 1965. This grew into a larger network called the Consultative Group for International Agricultural Research (CGIAR) now comprising 15 IARCs, including the International Centre for Agricultural Research in the Dry Areas (ICARDA) established in Syria in 1977, another key player in the international wheat and barley breeding network. Two of the major coordination responsibilities of CIMMYT are maintaining the world collection of wheat genetic resources – a public good protected by international treaty – and the facilitation of the International Wheat Nurseries.

After the initial impact of the Green Revolution in high production zones through exploitation of Rht-B1 and Rht-D1 dwarfing genes in conjunction with disease resistance, international breeding encompassed more challenging environments through, for example, international shuttle breeding between Brazil and Mexico to overcome problems associated with acid soils that restricted adoption of MVs. Another example is drought, which affects at least 30 million ha of wheat in the developing world. The approach focused initially on exploiting the inherent yield potential and disease resistance of MVs and later combined this with new stress-adaptive traits from wild wheat ancestors through wide crossing techniques. Adoption of modern varieties has increased substantially in drier areas between 1990 and 1997. In all environments, possibly the greatest threat to productivity is disease, especially those caused by fungal pathogens. International wheat breeding has placed great emphasis on genetic control of disease since resource-poor farmers generally lack the means to control diseases chemically.

The progestagen concentration of milk was measured in cows, at various times of the reproductive cycle by a competitive protein binding procedure and by radioimmunoassay. The apparent progesterone concentration was found to be significantly greater in milk (P <0–05) than in plasma, probably due to the presence of an unidentified metabolite(s). A simple, rapid radioimmunoassay was developed to measure directly the progestagen concentration in milk. A mean concentration of about 1 ng/ml at oestrus (or mating) increased, 20–24 days later, to 3–4 ng/ml in cows subsequently found to be non-pregnant and to about 7 ng/ml in pregnant animals. The possible application of this analytical procedure as the basis of a method of pregnancy diagnosis is discussed.

The principle of orthogonality in replicated experiments is discussed and the dangers of non-orthogonality emphasised. The modifications necessary in the ordinary procedure of the analysis of variance when applied to non-orthogonal data are developed, attention being paid to the shorter methods, applicable in certain cases, by which the heavy labour of computation necessary in the general method of fitting constants may be avoided.

Certain modifications in the design of replicated experiments, usually designated by the term confounding, are explained. The different types of confounding are discussed, together with their uses, and the appropriate methods of analysis are set out. The methods are applied to the analysis of an experiment on sugar beet (where a previous incorrect analysis is corrected) and an experiment on potatoes.

An indoor experiment involving 10 rumen-cannulated Romney sheep was conducted in May and June 1998 at AgResearch Grasslands, Palmerston North, New Zealand, under restricted feeding conditions, in order to test the hypothesis that animal factors, in particular rumen fractional outflow rate (FOR) and rumen volume, have an influence on the between-sheep variation in methane (CH4) emission. Sheep were fed 2-hourly on chaffed lucerne hay. Following an acclimatization period of 21 days, the experiment lasted 16 days. Energy and nitrogen (N) balances were measured on days 1–6. Cr-EDTA marker was continuously infused into the rumen from day 9 to 16, and rumen contents emptied and sampled on days 13 and 16. Particulate and fluid FOR were estimated using feed lignin and Cr-EDTA, respectively. Daily CH4 production was measured by the sulphur hexafluoride tracer technique on days 2, 5, 6, 12 and 15 of the experiment.

CH4 production (g/day) was positively correlated with the pool size of organic matter (OM) in the rumen (OM pool, g) (r=0·84, P=0·002), OM intake (OMI, g/day) (r=0·67, P=0·04), and the rumen fill (g, wet digesta) (r=0·76, P=0·01). Multiple regression analysis showed that CH4 production was best predicted (R2=0·88) as a function of OM pool and the molar % of butyrate; however, OM pool alone accounted for a large proportion (R2=0·71) of the variation in CH4 production.

CH4 yield (% gross energy intake, % GEI) was negatively correlated with the particulate FOR (%/h) (r=−0·75, P=0·01) and buffering capacity of rumen fluid (mmol HCl) (r=−0·72, P=0·02), but positively correlated with the digestibility of cellulose (r=0·66, P=0·04). Multiple regression analysis showed that CH4 yield was best predicted as a function of particulate FOR, OMI (g/kg liveweight0·75) and the molar % of butyrate (R2=0·88). Particulate FOR alone explained a large proportion (R2=0·57) of the variation in CH4 yield. Particulate FOR was negatively correlated with rumen fill (r=−0·69, P=0·03) and digestibility of cellulose (r=−0·65, P=0·04).

These results suggest that sheep with lower rumen particulate FOR (i.e. longer rumen retention times) had larger rumen fills and higher fibre digestibilities and CH4 yields. If rumen particulate FOR is to be used as a tool for CH4 mitigation, the repeatability of its relationship to CH4 emission must be assessed, preferably under grazing conditions.

Rumen biohydrogenation of dietary polyunsaturated fatty acids (PUFA) is a significant limitation on any attempt to manipulate the PUFA content of ruminant products (meat or milk). This study examined rumen biohydrogenation of PUFA, the effects of PUFA on other aspects of rumen metabolism and fatty acid flow to and digestion in the small intestine of steers fed on different sources of lipid. Animals were fed ad libitum on grass silage and one of four concentrates (60:40 forage:concentrate on a dry matter basis) containing differing sources of lipid: megalac (rich in C16:0; M), linseed (rich in C18:3n-3; L), fish oil (rich in C20:5n-3 and C22:6n-3; FO) and a mixture of linseed/fish oil (LFO). Diets were formulated so that total dietary oil intake was approximately 60 g/kg of the DM intake, approximately half of which was from the experimental test oil. Rumen NH3-N (P = 0·09) and total VFA concentrations (P = 0·007) were higher on L, FO and LFO compared to M. Dry matter intakes did not differ across treatments and averaged 7·2 kg/day. Intake and flow of fatty acids to the duodenum was 323, 438, 344 and 381 (S.E.M. 9·1; P < 0·001) and 432, 489, 412 and 465 (S.E.M. 18·5; P < 0·1) g/day for M, L, FO and LFO, respectively. Biohydrogenation of C18:1n-9 was lower than all the other unsaturated fatty acids and it was lower of FO and LFO compared to M and L, on average 66·1 and 72·2 %, respectively. Biohydrogenation of C18:2n-6 averaged 89·8 % across treatments and was lower (P < 0·05) on M compared to L and FO. Biohydrogenation of C18:3n-3 averaged 92·1 % across treatments and was lowest on M (88·8 %) and highest on L and LFO (94·3 %). Biohydrogenation of C20:5n-3 and C22:6n-3 averaged approximately 91 and 89 % across the treatments, respectively. Small intestinal digestibilities of all fatty acids were high. In conclusion, feeding different sources of lipid with different fatty acid composition had significant effects on rumen function. The PUFA in whole linseed were only partially protected from biohydrogenation by the seed coat and in contrast to previous reports the C20 PUFA in fish oil were biohydrogenated to a large extent.

A field and a glasshouse experiment were carried out to assess in winter wheat genotypes the interrelationships between stem and ear development, and to relate these to number of grains and yield. Capelle-Desprez (gai/rht2) and Hobbit ‘s’ (Gai/Rht2) were used together with four F4 lines from the cross between them. Two of the lines were homozygous for gai/rht2 and two lines homozygous for Gai/Rht2. Within each pair of homozygous lines there was a contrast between a tall and a short genotype.

There were no consistent differences in shoot apex morphogenesis between the genotypes, and, in general, similar numbers of spikelets and floret primordia were initiated in all genotypes. This contrasts with the hypothesis that genotypes based on Norin 10 (Gai/Rht2) have a fundamentally different pattern of ear morphogenesis. The main differences between genotypes were in dry-matter partitioning to stem and ear prior to anthesis and these were associated more consistently with the dwarfing gene than with height per se. Gai/Rht2 genotypes partitioned more dry matter to the ear during its development and less to the stem and this was reflected in greater ear weight at anthesis, more fertile florets and grains per spikelet, a higher harvest index and higher grain yield than the gai/rht2 genotypes. Although similar numbers of floret primordia were initiated in all genotypes, assimilate partitioning during development determined the proportion which developed into fertile florets at anthesis.

1. A study has been made over two growing seasons of the volatile fatty acid production from ley grass obtained from a single area of one field, under the action of rumen liquor in vitro.

2. The acida estimated were acetic, propionic, butyric, valeric and caproic. On only one occasion did the last named appear in measurable amount.

3. The main acid produced in early stages of the year is acetic acid, but as the season advances propionic acid becomes the acid of major production.

4. In parallel cellulose runs it was found that propionic acid was invariably produced in greater amount than any other acid.

5. An attempt has been made to eliminate, by using cellulose as a standard, the effects of using different rumen liquor samples.

6. The dried samples, corresponding to the fresh material, were invariably found to yield acetic acid in greater proportion than propionic acid. The average percentages of acids obtained from the dried grass were similar to those found by other workers. It is suggested that the variations between the fresh and dried grass results are due to changes in carbohydrate content resultant upon storage of the latter.

Sheep prepared either with a rumen cannula, or with a rumen cannula plus re-entrant cannulae in the duodenum and ileum were fed fresh ‘Ruanui’ perennial ryegrass (R), ‘Manawa’ short rotation ryegrass (M) and white clover (C) at dry-matter intakes ranging from 450 to 1000 g/24 h. Paper impregnated with chromic oxide was given once daily via the rumen fistula as a marker.

Amounts of non-ammonia-nitrogen (NAN) entering and leaving the small intestine and nitrogen (N) excreted in the faeces were regressed against the intake of N for each sheep and these equations (all significant P < 0·05) were used to calculate the extent of digestion in the stomach, the small intestine and the large intestine at two levels of OM intake (500 and 800 g daily). Amino acid analyses of feed and digesta samples are also presented together with calculations of the apparent absorptions of individual amino acids from the small intestine.

At an intake of 500 g OM amounts of NAN entering the small intestine were significantly greater (P < 0·05) in sheep given M than in sheep given R or C. At an intake of 800 g OM amounts in sheep given M were significantly greater (P < 0·01) than in sheep given R. NAN leaving the small intestine and N excreted in the faeces were similar for all three diets.

At an intake of 800 g OM apparent absorptions of NAN from the small intestine of sheep given M (0·47 × N intake + 2·8 g/24 h) and sheep given C (0·51 × N intake – 1·3 g/24 h) were significantly greater (P < 0·01 and P < 0·05 respectively) than of sheep given R (0·41 × N intake – 0·5 g/24 h). As intake of herbage was increased the partition of digestion altered.

Only small differences between herbages were found in the amino acid composition (g amino acid/100 g protein) of either duodenal or ileal digesta, but because of the large differences in the flows of NAN, the apparent absorptions of individual amino acids from the small intestine were much higher in sheep given M than in sheep given the other two species.

The results are discussed in the light of available information on sites of digestion of herbage diets in sheep.

Three techniques, lamb-suckling and test-weighing, hand- or machine-milking following administration of oxytocin, and measurements of body-water turnover, used for the estimation of milk yield in suckled sheep are discussed in relation to sources of error and comparability of estimates.

Results are presented from an experiment in which the lamb-suckling and the oxytocin techniques were compared. The mean daily milk production of Scottish Blackface and East Friesland × Scottish Blackface ewes suckling single and twin lambs was measured by both methods at weekly intervals for 6 weeks. All ewes were well nourished to ensure full expression of milking capacity. Additional information from a similar experiment on Merino ewes is also presented.

The oxytocin method gave higher estimates of mean daily milk yield in the first week of lactation, especially in the groups of ewes rearing single lambs and in the Blackface breed. By the second week in most groups and the third week in all groups there were no significant differences between estimates made by the two methods and the relative values of the estimates were not affected by level of milk production, number of lambs suckled or genotype of ewe and lamb.

The relation between the ratio of the natural 12C and 13C isotopes of carbon in the feed and resultant faeces of animals was studied to develop a technique for estimating the proportion of C3 species (tropical legumes) and C4 species (tropical grasses) selected by grazing animals.

In general, theδ13C values (see text for definition) of faeces from rabbits, sheep, goats and cattle were lower (more negative) than those of the corresponding feeds by from 0·4 to 2·0. This was possibly due to contamination in the gut by tissues or fluids with lower δ13C values. When C4 and C3 feeds were alternated, cattle took about a week to fully achieve the new level (δ13C of – 28·7 on the C3 feed and – 13·1 on the C4 feed) in the faeces. This time lag is associated with the time taken for the feed to move through the digestive tract.

When mixed C3 and C4 feeds were fed to rabbits, sheep, goats and cattle there was a negative linear relation between percentage legume (C3) in the feed and the δ13C of the faeces (P < 0·01). A decrease in one unit in the δ13C value was associated with an increase of 7·0–8·5% legume in the diet.

Estimation of the percentage legume in the feed from the δ13C value of the faeces and of the C3 and C4 components of the diet, resulted in a consistent over estimation of the legume component because the faeces had lower values than the corresponding feeds. This bias was removed if the prediction was based on the δ13C of the feeds minus 1 unit; the legume percentage in the diets of the sheep, goats and cattle could then be estimated with a precision of about ± 5%.

Differences in digestibility between the C3 and C4 components greatly bias the estimations. This bias in the diets fed to rabbits was effectively removed by using in vitro organic matter digestibility values of the two components to correct for the differences. Legume percentage in the diet could then be estimated with a RSD of ± 3%.

Advantages and disadvantages compared with alternative methods of estimating the diet of grazing animals are discussed.

Much has been written and said about genotype-environment (GE) interactions and the particular problems which they pose for plant breeders. It is not the purpose of this article to dwell upon every aspect of this story, but rather to discuss how these problems came to be recognized, to comment upon the various techniques which have been employed in seeking a solution to them and to suggest what developments might lie ahead.

Ten dairy cows were allocated into three groups according to milk productivity (four high, four intermediate and two dry cows, respectively). Heat production and heart rate, but not rectal temperature, were significantly different (P < 0·05) between groups. Heat production increased during feeding in the morning and in the afternoon and reached a peak 3 h later. Minimum heat production was observed in the early morning before feeding. The diurnal pattern for heart rate reflected that of heat production. These results suggest that cooling dairy cows during hot summer days is most effective at feeding time and 3 h afterwards.

A grazing experiment was conducted for 8 weeks in the spring/summer of 1993 at Palmerston North, New Zealand, to study the effects of condensed tannins (CT) in Lotus corniculatus (birdsfoot trefoil cv. Grasslands Goldie) upon the lactation performance of ewes rearing twin lambs. Effects of CT were evaluated by studying the responses of ewes to twice daily oral supplementation with polyethylene glycol (PEG; MW 3500), which binds and inactivates CT. A rotational grazing system with restricted feed allowance was used. Measurements were made of pre- and post-grazing herbage mass, the composition of the feed on offer and diet selected, voluntary feed intake (VF1), milk yield and composition, liveweight gain and wool production. The concentration of metabolites in rumen fluid and in blood plasma was also measured. Lotus contained 35–5 g total nitrogen and 44–5 g total CT/kg dry matter in the diet selected, with an in vitro digestibility of 73%. At peak lactation (weeks 3 and 4) milk yield and composition were similar for control (CT-acting) and PEG-supplemented (CTinactivated) ewes but, as lactation progressed, the decline in milk production and in the secretion rates of protein and lactose were less for control than for PEG-supplemented ewes. In mid and late lactation (weeks 6–11), control ewes secreted more milk (21 %), more milk protein (14%) and more lactose (12%) than PEG-supplemented ewes. Milk fat percentage was lower for control than for PEG-supplemented ewes, but secretion rates of fat were similar for the two groups. VFI, liveweight gain and wool growth were similar for both groups. Plasma urea and glucose concentrations were lower for control than for PEG-supplemented ewes, but concentrations of non-esterified fatty acids (NEFA), growth hormone and insulin were similar for the two groups. The concentrations of ammonia and molar proportions of iso-butyric, iso-and n-valeric acids in rumen fluid were lower for control than for PEG-supplemented ewes; molar proportions of acetic, propionic and H-butyric acids were similar for the two groups. It was concluded that for ewes rearing twin lambs grazing L. corniculatus, the action of CT increased milk yield and the secretion rates of protein and lactose without affecting VFI, thereby increasing the efficiency of milk production. The increased milk production did not appear to be mediated by effects on plasma concentrations of growth hormone or insulin.

Five experiments, involving 37 varieties, were carried out over three seasons to estimate the increase in yield potential in spring barley achieved by plant breeding during the last 100 years. Most of the important spring barley varieties grown in Britain between 1880 and 1980 were represented. In three experiments disease was controlled by a fungicide while in the other two experiments fungicide application was a main treatment. To prevent yield loss due to lodging, plants were supported as a main treatment in two trials.

In all experiments most of the modern varieties yielded more than the older ones. The genetic gain in yield was 0·39% per year during the 100-year period and 0·84% per year between 1953 and 1980.

Modern varieties had higher grain yields, shorter straw, and higher harvest index denned as the proportion of grain dry weight to total above-ground dry weight; more of the tillers they produced survived to give ears. There was a weak association between biological yield (total above-ground dry weight) and grain yield.

It is suggested that although much of the improvement in yield described here could be attributed to increased harvest index, the scope for further improvement in this character may be limited. Further yield improvements might be achieved by combining high biological yield with high harvest index.

An agro-economic simulation model was developed to facilitate comparison of the impact of management, market and biological factors on the cost of providing ruminant livestock with feed grown on the farm (home produced feed). Unpredictable year-to-year variation in crop yields and input prices were identified as quantifiable measures of risk affecting feed cost. Stochastic analysis was used to study the impact of yield and input price risk on the variability of feed cost for eight feeds grown in Ireland over a 10-year period. Intensively grazed perennial ryegrass was found to be the lowest cost feed in the current analysis (mean cost €74/1000 Unité Fourragère Viande (UFV)). Yield risk was identified as the greatest single factor affecting feed cost variability. At mean prices and yields, purchased rolled barley was found to be 3% less costly than home-produced spring-sown barley. However, home-produced spring barley was marginally less risky than purchased barley (coefficient of variation (CV) 0·063 v. 0·064). Feed crops incurring the greatest proportion of fixed costs and area-dependent variable costs, including bunker grass silage, were the most sensitive to yield fluctuations. The most energy input-intensive feed crops, such as grass silage, both baled and bunker ensiled, were deemed most susceptible to input price fluctuations. Maize silage was the most risky feed crop (CV 0·195), with potential to be both the cheapest and the most expensive conserved feed.

In a preliminary experiment, 74 faba bean genotypes including winter genotypes (autumn-sown) and spring genotypes (spring-sown) and isogenic population pairs (tannin-containing v. tannin-free and vicine/convicine-high v. vicine/convicine-low), were analysed for the chemical composition of their seeds. A large variability was found for the main constituents (starch, protein and fibre). Autumn-sown genotypes contained 2·3% less proteins but 2·5% more starch in the seed dry matter (DM) than spring-sown genotypes. The vc− gene, which lowers the vicine and convicine contents, did not significantly modify the main seed components in the isogenic comparisons. The zt1 and zt2 genes, which eliminate condensed tannins in the seed coats, lowered by 2·1% the proportion of the seed coat in the DM. In the isogenic comparisons, the zt2 gene had a stronger effect than zt1 in reducing the total seed fibre and increasing the protein content.

In a second experiment, from the original 74 genotypes, 12 contrasted genotypes were selected and multiplied for animal nutrition trials. Their chemical analysis confirmed the variability between the faba bean categories observed in Expt 1, but detailed chemical analyses illustrated the variability in amino acid, fatty acid, amylose and oligosaccharide composition, trypsin inhibitory activity, condensed tannins, lectins and phytic phosphorus contents.

Milk urea nitrogen (MUN) concentration in dairy cows may serve as an on-farm indicator to guide nutritional strategies and to help reduce emissions of nitrogen (N) to the environment. Excretion of urinary urea nitrogen (UUN) is positively related to MUN, but the relationship is highly variable. The accuracy of MUN as a predictor of UUN may improve when various factors that affect this relationship can be taken into account. The current review discusses the impact of a number of UUN : MUN ratio influencing factors related to: physiological mechanisms in the dairy cow, farm management, differences between individual cows, nutrition and analysis methods for MUN. Factors related to variation in water intake, urine production, dietary protein level, body weight (BW) and time and frequency of feeding and milking are shown to affect MUN and its relationship with UUN. In addition, a number of factors are discussed that are likely to affect this relationship such as biological rhythm, renal reabsorption of urea during periods of protein deficiency and breeding value for MUN. Accounting for these above-mentioned factors in the relationship between MUN and UUN might substantially improve the applicability and accuracy of MUN as a predictor of protein utilization efficiency and UUN.

In this paper we present a series of resolvable incomplete block designs suitable for variety trials with any number of varieties v in the range 20 ≤v ≤ 100. These designs usefully supplement existing square and rectangular lattices. They are not necessarily optimal in the sense of having smallest possible variances but their efficiencies are known to be high.

1. An analysis has been made of the production records of over 1400 heifers got by A.I. bulls at five non-Board centres. The A.I. heifers, when compared with other heifers in the same herd in the same year, showed no significant increase in yield but a significant increase in fat content.

2. The difference in yield between A.I. and contemporary non A.I. heifers was not significantly different in herds at all levels of production. This would suggest that only a small proportion of the differences in yield between the herds in the sample was genetic in origin. In fat percentage, there was some evidence, though not very conclusive, that rather more of the inter-herd variation was genetic.

3. Significant differences were found between the progeny groups by different sires in both milk yield and fat content. The standard deviation of the variation due to the breeding value of the sire was 37 gal. and 0·13% fat respectively.

4. Methods of interpreting progeny tests made under A.I. conditions are discussed theoretically. It is concluded that the comparison with heifer contemporaries in the same herd is most satisfactory for milk yield and the simple average is best for fat content.

5. The relevance of these results to the genetic structure of breeds is discussed. They are in good agreement with the conclusions from pedigree analysis that the genetic differences between the dominant group of herds and the mass of commercial cattle in each breed is not great.