Male rats were fed on either a non-purified rodent diet (JS) or cholesterol-free purified diets containing wheat bran (WB) or oat bran (OB). Some animals were allowed normal access to their faeces for coprophagy (coprophagy +), while in others coprophagy was prevented by placement of a plastic cup over the anus (coprophagy −). Direct ingestion of faeces from the anus was observed in the former groups. Food intake was unaffected by diet or coprophagy status and body weight gain was unchanged with OB − and JS − but was significantly lower with WB −. Plasma cholesterol was highest with WB and equally lower with OB and JS and was unaffected by coprophagy status. Plasma triacylglycerols were highest with OB and were unaffected by coprophagy status. Caecal digesta mass was highest with JS, intermediate with OB and lowest with WB. Digesta mass was unaffected by coprophagy status with WB and JS but was higher with OB −. Digesta moisture content was lowest with WB + but highest with WB −. Digesta volatile fatty acid (VFA) concentrations were similarly lower with OB + and OB −, but were significantly lower with JS − and WB − than in the corresponding coprophagy + group. In all groups digesta butyrate concentrations were reduced by coprophagy prevention. Pools of total VFA, acetate and butyrate in the digesta were highest with JS. Pools of total VFA in digesta were highest with JS +, OB + and OB − and lowest with WB + and WB −. The propionate pool was highest with OB −, intermediate with OB +, and equally low in all other groups. The pool of butyrate was highest with JS + and lowest with OB −. Effects of oats and wheat on plasma cholesterol in the rat do not seem to be mediated through faecal re-ingestion.

The present study was designed to assess the feasibility of using small doses of a stable isotope of Zn to follow plasma kinetics over a 10 d period and, hence, make deductions about Zn turnover and body pool sizes. At the beginning of the 10 d metabolic balance, two adults, consuming their habitual diet, were given an intravenous injection of 70Zn. There was a fourfold difference in the administered dose between the two subjects (0·445 and 2·078 mg). Blood samples were taken at regular intervals and plasma enrichment with 70Zn measured by thermal ionization mass spectrometry. Urine and faeces were collected and analysed for Zn and 70Zn. Kinetic analysis of the plasma 70Zn decay by several different methods was undertaken. It was apparent from both deconvolution analysis of the short-term (0–90 min) decay data and four-compartment modelling of the longer-term (0–24 h) data that isotopic Zn very rapidly equilibrates with the plasma Zn and with a rapidly exchanging non-plasma pool, probably located within the liver. This latter pool appears to contain less than 10 mg Zn and the peak of isotope enrichment occurs at about 20 min post injection. The later decay of plasma Zn enrichment appears to be dictated by exchange with a much larger pool of approximate size 350 mg.

This study evaluated Ca intake in Hong Kong Chinese children and examined the association between long-term Ca intake and bone mineral content (BMC) in children. Longitudinal dietary intake from birth to 5 years was obtained in 128 children (sixty-seven boys, sixty-one girls). Ca intakes were evaluated by dietary history and cross-checked with food frequency and 24 h recall. At age 5 years BMC was determined at the one-third distal radius of the right arm using single-photon absorptiometry. The mean Ca intake of 133 children at 5 years was 546 (SD 325) mg/d. Milk was the chief source of Ca (43·5 %). From birth to 5 years, 90 % of the children had been taking milk regularly. The mean BMC and bone width (BW) of these children were 0·317 (SD 0·042) g/cm and 0·756 (SD 0·074) cm respectively. BMC was not correlated with current intakes of Ca, energy and protein but was positively correlated with weight (r 0·57), height (r 0·47) and BW (r 0·66). However, cumulative Ca intake throughout the past 5 years showed significant correlation with BMC (r 0·235, P = 0·0133). The significant correlation remained even after weight, height, BW, sex. and cumulative intakes of energy and protein were adjusted in multiple regression analysis (r 0·248, P = 0·0107). Moreover, using principal component analysis, Ca intake during the 2nd year of life had the strongest correlation with BMC at 5 years (r 0·240, P = 0·02). Ca intake of Hong Kong Chinese children was higher than the RDA of the Food and Agriculture Organization/World Health Organization (1962) and achieved 66% of the current US recommendation (National Research Council, 1989). The increased regular milk consumption reflects a significant change in dietary habits of the younger generation. Children with a habitually higher Ca intake throughout the past 5 years, particularly in the 2nd year, were found to have higher BMC.

Two experiments were conducted to determine the available P requirements of grower and grower/finisher pigs and to define the conditions for conducting a growth assay for P availability. In the first experiment, diets with four levels of calculated available P (1–4 g/kg) and four Ca: available P ratios (1·7–2·9) were used to determine the available P requirements of grower pigs. The diets were formulated by substituting the required amounts of limestone and sodium tripolyphosphate for sugar in a soya-bean meal and sugar-based diet. In addition to measuring growth responses, a range of bones were examined to determine the most suitable criteria for assessing the response to available P. There was a small quadratic response of feed intake and growth rate of the pigs to level of available P, with maximum responses occurring to approximately 3 g available P/kg (P < 0·05). There were linear depressing effects of increasing Ca:available P ratios on carcass gain and feed conversion ratio (P < 0·01) but most of these effects occurred when the ratio exceeded 2·5:1. All bone variables examined increased linearly (P < 0·05) or curvilinearly (P < 0·01) with increasing available P concentration. In general, these variables were not affected by the Ca: available P ratio. The results of the growth responses and bone development indicate that the grower pig requires approximately 3 g available P/kg. However, for availability assays, where linearity of response is needed, the dietary concentration of available P should be a maximum of approximately 2 g/kg. In the second experiment four levels of calculated available P (1–4 g/kg) with a Ca: available P ratio of 2·5:1 were used to determine the available P requirements of grower/finisher pigs from 20 to 90 kg live weight. At 50 kg live weight the dietary available P concentration for half the pigs fed at 2, 3 and 4 g available P/kg was reduced to 1, 2 and 3 g/kg respectively. The pigs were fed ad lib. and growth performance, bone characteristics, P retention and ash concentration in the empty body were taken as response criteria to assess P adequacy. Among the variables tested, the ash concentration in the radius/ulna bone and P and ash concentrations in the empty body appeared to be more responsive than other variables to the changes in dietary P levels. Based on these variables, the P requirements for growth and bone development of growing pigs from 20 to 50 kg live weight was 3 g/kg and reduced to 2 g/kg for finisher pigs from 50 to 90 kg live weight.

Experiments were conducted with pigs and rats to determine the availability of P in feeds. Initially, the availability of P in a soya-bean meal and field peas (Pisum sativum cultivar Early Dun) was assessed using a slope–ratio assay for grower pigs. Three different levels of either monosodium phosphate (MSP), soya-bean meal or field peas were added to a basal sucrose–soya-bean meal diet (2·5 g/kg P) to give three levels of P (3·0, 3·5 and 4 g/kg) for each source. The diets were offered for 35 d at three times maintenance energy requirements to female pigs initially weighing 20 kg live weight. Several bone variables and the ash and P concentration and retention levels in the empty body were used as criteria of availability. The responses to MSP were linear for all variables. However, responses to P in the test proteins, particularly soya-bean meal, were mostly non-linear, except for ash and P concentrations and retentions in the empty body. The estimates of P availability in the soya-bean meal and in the field peas were dependent on the criteria used to assess availability. Using bone variables as the criterion, the availabilities of P in soya-bean meal and field peas were approximately 0·17 and 0·38 respectively. Using ash or P concentrations or retentions in the empty body as the response criterion, the availabilities of P in soya-bean meal and field peas were 0·61 and 0·38 respectively. The mean retention values for P from MSP, soya-bean meal and field peas were 0·74, 0·33 and 0·21 respectively. Experiments were then conducted to define the conditions necessary for a slope–ratio assay for P availability with growing rats to determine if they could be used as a model to assess P availability for pigs. These experiments examined level of P in the diet, the effect of Ca:P ratio and criteria for assessing response (growth rate, feed intake, feed conversion ratio, bone ash and bone bending moment). The results indicated that a suitable range of dietary P for a slope–ratio assay in rats was 1·8–3·5 g/kg, with Ca: P ratio of between 1·3 and 6·2, and that ash content or bone bending moment of the femur bone were suitable criteria of response. However, using bone bending moment as the criterion of response, the availability of P in soya- bean meal was 0·81, which was considerably higher than the estimate with pigs. Overall the results indicated that the estimates of the availability of P in feeds for pigs were dependent on the criteria used to assess availability and that the value for rats led to an overestimate of the value for pigs.

Two experiments were conducted (1) to determine the effects of phytase (EC 3.1.3.26) on the digestibility and availability of P in soya-bean meal for growing pigs and (2) to compare growth v. digestibility variables for assessing the availability of P. In the first experiment the effect of phytase on P availability was assessed in a growth assay using a slope–ratio design of treatments. Two different levels of either monosodium phosphate (MSP) or soya-bean meal were added to a basal sugar–soya-bean-meal diet (2·5 g P/kg) to give two levels of P (g/kg): 3·25 and 4·0 for each source. An additional five diets were supplemented with phytase. The ten diets were offered ad lib. for 35 d to female pigs initially weighing 20 kg live weight. In addition, the relative effectiveness of different variables for assessing P availability were compared: bone bending moment, ash in various bones, and ash and P in the empty body. The addition of phytase increased growth rate (g/d) (741 v. 835; P < 0·05), lowered the food conversion ratio (2·37 v. 2·16; P < 0·01), and increased protein deposition (g/d) (108 v. 123; P < 0·05), protein retention (kg/kg) (0·33 v. 0·36; P < 0·05), energy retention (MJ gross energy/MJ digestible energy) (0·36 v. 0·38; P < 0·05) and the availability of P in soya-bean meal from 0·11 to 0·69 when bone bending moment was the criterion of availability. All other criteria for assessing availability were unsuitable. In the second experiment the availability of (P) in soya-bean meal was assessed in a digestibility experiment with grower pigs using diets 1–5 as for Expt 1 arranged in a slope–ratio design of treatments. In addition, the effects of phytase supplementation on the apparent digestibility of P, dry matter, crude protein (N × 6·25) and energy were determined. The diets were offered at three times maintenance energy requirements to male pigs initially weighing approximately 30 kg live weight and total collection of faeces was conducted over a 10 d period. The availability of P in the soya-bean meal was 0·66 using digestible P intake as the criterion of response. The apparent digestibility of P in soya-bean meal was 0·42. Phytase supplementation increased the apparent digestibility of soya-bean meal P to 0·69 (P < 0·01) but had no effect on the faecal digestibility of dry matter or crude protein. Overall these experiments indicate that (1) estimates of P digestibility and availability were unlikely to be interchangeable and (2) phytase was effective in releasing much of the bound P in soya-bean meal.

Incorporation of N-acetylglucosamine-specific agglutinins from wheat germ (Triticum aestivum; WGA), thorn apple (Datura stramonium) or nettle (Urtica dioica) rhizomes in the diet at the level of 7 g/kg reduced the apparent digestibility and utilization of dietary proteins and the growth of rats, with WGA being the most damaging. As a result of their binding and endocytosis by the epithelial cells of the small intestine, all three lectins were growth factors for the gut and interfered with its metabolism and function to varying degrees. WGA was particularly effective; it induced extensive polyamine-dependent hyperplastic and hypertrophic growth of the small bowel by increasing its content of proteins, RNA and DNA. Furthermore, an appreciable portion of the endocytosed WGA was transported across the gut wall into the systemic circulation, where it was deposited in the walls of the blood and lymphatic vessels. WGA also induced the hypertrophic growth of the pancreas and caused thymus atrophy. Although the transfer of the gene of WGA into crop plants has been advocated to increase their insect resistance, as the presence of this lectin in the diet may harm higher animals at the concentrations required to be effective against most pests, its use in plants as natural insecticide is not without health risks for man.

Gnotobiotic growing rats harbouring either a whole human faecal flora or single human strains of Escherichia coli (EM0) or Bacteroides vulgatus (BV8H1) were fed for 7 weeks on semi-synthetic diets in which the protein source was either soya-bean meal (SM) or rape-seed meal (RM). For each bacterial status the RM-diet group was compared with the control group fed on the SM diet. The association of human faecal flora with the RM diet was responsible for reduced feed intake and reduced weight gain, an enlargement of the liver and thyroid and a decrease in both thyroxine and triiodothyronine plasma levels. The association of the B. vulgatus BV8H1 strain with the RM diet reproduced all these effects, except that triiodothyronine plasma levels were not significantly modified. Rats inoculated with the E. coli EM0 strain and fed on the RM diet exhibited a goitre and lowered thyroxine and triiodothyronine plasma levels. These results show that the human intestinal microflora may be involved in glucosinolate metabolism when cruciferous vegetables are consumed by man. The specificity of the symptoms observed according to the rat bacterial status supports the hypothesis that bacteria yield specific toxic glucosinolate derivatives according to their enzymic potential.

Pancreatic and intestinal growth rates were measured in mice fed on raw soya-bean flour (RSF) for up to 24 weeks. Control animals were fed on standard chow. The effects of RSF on the mouse pancreas resembled that seen in rats, showing hypertrophy with some hyperplasia. A marked increase in small intestinal weight was also found in mice fed on RSF but not in rats fed on this diet. Histological studies showed an increase in both villous and crypt thicknesses in the small intestine from these mice, and DNA, RNA and protein measurements indicated that the increase in intestinal weight was due to hypertrophy and hyperplasia of the mucosal layer. To determine whether the intestinal growth in mice fed on RSF was purely a response to the trypsin inhibitor (TI) component of the diet, pancreatic and intestinal growth rates were also determined in mice fed on the synthetic trypsin inhibitor camostate, at levels of 0·5 or 2 g/kg in rat chow, for periods of 1–8 weeks. Control animals were fed on standard chow. RSF and 0·5 g camostate/kg had similar trypsin inhibitor activities (measured against bovine trypsin), and both caused similar increases in pancreatic weight, DNA, RNA and protein content. However, 0·5 g camostate/kg did not affect small intestinal weight. Chow containing 2 g camostate/kg contained twice as much TI activity as the RSF diet but produced only a small increase in small intestinal weight at 2 and 8 weeks. This intestinal growth was significantly less than that seen with RSF. The present study shows that, in the mouse, RSF or a diet containing camostate in the appropriate dose produces pancreatic growth comparable to that seen in the rat. RSF also causes intestinal growth, but camostate-containing diets have little or no effect on the growth of the intestine.

Germ-free growing rats were fed on a glucosinolate-rich diet (rapeseed-meal-based) and compared with counterparts fed on a glucosinolate-free diet (soya-bean-meal-based), both diets being isonitrogenous and isoenergetic. For each diet half the animals received phenobarbital in drinking water as an inducer of xenobiotic-metabolizing enzymes. Some of the usual deleterious glucosinolate-linked effects, i.e. kidney hypertrophy and reduction in growth and feed intake, were followed and three of the major hepatic xenobiotic-metabolizing enzymes were investigated. Growth rate, dietary intake and kidney weight were not altered by glucosinolates in the absence of intestinal microflora, whether the animals were treated with phenobarbital or not. As far as the hepatic xenobiotic-metabolizing enzymes are concerned, the specific level of cytochrome P450 and the specific activities of glutathione-S-transferase (EC 2.5.1.18) and UDPglucuronosyltransferase (EC 2.4.1.17) remained unchanged in rats receiving the glucosinolate-rich diet compared with the control animals. Despite the low dose given, phenobarbital displayed its usual inducing effect on all three enzymes, similar whatever the diet. A previous counterpart experiment performed with conventional animals had shown that glucosinolate feeding led to large alterations of the variables herein studied, some of these modifications being hugely enhanced by a phenobarbital treatment. Therefore, the present results obtained on germ-free animals prove that alterations of the xenobiotic-metabolizing enzymes induced by glucosinolates are somehow mediated by the intestinal microflora. Furthermore, the involvement of those enzymes in glucosinolate toxicity definitely requires the presence of the intestinal microflora.

Cotyledons and hulls were prepared from twelve varieties of field beans (Vicia faba L.). Adult cockerels were tube-fed either beans, cotyledons or hull diets containing high or low levels of protein. Metabolizable energy coefficients and starch digestibility coefficients were determined for beans, cotyledons and hull diets. Lipid digestibility coefficients from hull diets were also determined. When cotyledons were fed there were no significant differences in the way in which adult cockerels metabolized energy or digested starch from the proanthocyanidin-free and proanthocyanidin-rich varieties (0·780, 0·908, 0·775 and 0·918 respectively). When beans were fed, however, both energy metabolizability and starch digestibility decreased due to the presence of hulls, with proanthocyanidin-rich hulls decreasing values the most to 0·660 and 0·819 respectively, and proanthocyanidin-free hulls decreasing values to a lesser extent to 0·709 and 0·886 respectively. Diets composed of proanthocyanidin-rich hulls depressed metabolizable energy and maize starch digestibility. Their effect on maize starch digestibility, however, was considerably less than that on bean starch. Lipid digestibility was enhanced by proanthocyanidins but only when the protein content of the diet was high. There was a significant correlation (P < 0·05) between the vanillin and anthocyanidin formation methods for the estimation of proanthocyanidins (r 0·779). There was also a highly significant regression of bean starch digestibility v. proanthocyanidin content of coloured-flowered bean hulls (P < 0·001). The regression of maize starch digestibility v. hull proanthocyanidins was also significant at P < 0·005.

Deterministic mathematical equations are derived to describe the pattern of marker excretion in the faeces of ruminants under steady-state conditions when diffusion and viscosity concepts are introduced into a simple two-compartment scheme of the gastrointestinal tract. The basic scheme comprises a pure-mixing pool obeying first-order kinetics and a second compartment exhibiting streamline flow. Introduction of a velocity gradient, longitudinal diffusion or both into the second compartment, even with various simplifying assumptions, yields analytically insoluble equations. The impact of these mechanisms is to be investigated numerically rather than analytically in future work.