Three successive sowings of marrow-stemmed kale (cv. Maris Kestrel) were grazed with lambs (initial weight 22·5 kg) for two consecutive 6-week periods, so that crop yield at the time of grazing was kept relatively constant at 6·5 t D.M./ha. Three crop allowances, 3·9 (H), 2·6 (M) and 1·3 (L) kg D.M./animal/day were imposed in a replicated design. Halfof the 32 lambs grazing each allowance received intraperitoneal injections of L-methionine + L-threonine three times per week. Twelve lambs were slaughtered at the start of the experiment and 48 at the end of each period.

Leaf tissue, the diet of the H groups, contained 0·5% D.M. as S-methylcysteine sulphoxide (SMCO) and 0·7–1·5% total N as nitrate. Whole plants, the diet of the L groups, contained 0·75% D.M. as SMCO and 4·0–7·5% total N asnitrate. The diet of the M groups was intermediate.

Rates of live-weight gain (LWG), empty-body weight gain (EBWG) and carcassweight gain (CWG) were lower in period 1 than in period 2 for animals grazing at all three allowances. A Heinz-Ehrlich body anaemia developed that was more severe in the first period, but did not differ between groups receiving the different allowances. Methaemoglobin was only detected in significant quantities at the time of peak Heinz body formation in period 1. CWG/EBWG was increased by the amino acid supplementation during period 1 but not period 2, showing that there was a temporary amino acid deficiency corresponding to the time when tho anaemia was most severe and body growth most depressed.

In both periods animals grazing at L allowance had lower rates of LWG, EBWG and CWG than M and H animals, who performed similarly. Most of the differences in EBWG between feed allowance treatments and between periods could be explained by changes in estimated intake. Wool growth rates increased with increasing feed allowance and were slightly stimulated by amino acid supplementation.

Copper true availability was calculated to be 1·6–2·7% with the lowest values for leaf tissue due to its high sulphur content. Truly available Cu/kg plant D.M. was below the minimum requirement for lamb growth, thus leading to a reduction in total liver Cu as the experiment progressed. Some indication was obtained that this could have been depressing LWG in the second period, especially in H lambs. Slight increases in thyroid gland weight were observed as the experiment progressed, but these were considered unlikely to have depressed body growth. It was concluded that the major factor limiting the feeding value of B. oleracea was its high content of nonprotein sulphur compounds, especially SMCO.

Animals fed sole diets of kale (Brassica oleracea) were compared with animals fed ryegrass-clover pasture grown on the same soil type in two experiments. In Expt 1 young cattle grazed the two forages for 24 weeks, with supplementary copper and iodine being administered by injection. In Expt 2 young sheep were individually fed the two forages indoors at equal D.m. intake.

Animals grazing kale in Expt 1 showed the characteristic symptoms of haemolytio anaemia from ruminal fermentation of S-methyl cysteine sulphoxide (SMCO) (Smith, 1974). This was most severe over the first 6 weeks, during which live-weight gains were very low (250 g/day). In the absence of copper supplementation animals grazing kale showed symptoms of copper deficiency. This was characterized by live-weight gain remaining low throughout the experiment (mean 280 g/day), rapid depletion of liver copper reserves, progressive reductions in serum copper concentration, reductions in erythrocyte copper and reduced glutathione (GSH) concentrations and a massive hepatic accumulation of iron. Copper deficiency only slightly lowered heart muscle copper concentration in kale-fed cattle, and this was counteracted byheart hypertrophy. The major effects of copper deficiency in kale-fed cattle were in erythrocytes, and a metabolic diagram is presented showing these effects to be biochemically similar to those produced by ruminal fermentation of SMCO.

Copper supplementation of animals grazing kale increased live-weight gain (mean 425 g/day), reduced Heinz body formation, allowed the animals to recover gradually from the haemolytic anaemia and prevented other symptoms of copper deficiency. In contrast, animals grazing ryegrass-clover pasture showed only a very mild depletion of copper, there being no response in live-weight gain to copper supplementation.

Activity of the enzyme glutathione peroxidase (GSH-Px) in whole blood was dependent upon blood selenium concentration in cattle fed both diets. In cattle fed on kale, bub not on pasture, reductions in erythrocyte GSH due to ruminal fermentation of SMCO and to copper deficiency were also associated with depressed blood selenium status.

Glucosinolates were present in the kale (11μM/g D.M.) but absent from the pasture diet. Despite this, neither T4 production from the thyroid gland nor the conversion of T4 to T3 appeared to be impaired by kale feeding in either Expt 1 or Expt 2. In Expt 1 serum T3 concentration was better relatedto live-weight gain than was serum T4 concentration, in accord with T3 being the active form of the thyroid hormone.

It is concluded that supplementation with copper but not iodine is essential where growing cattle are fed sole diets of kale for periods in excess of 12 weeks

In Expt 1 40 lactating British Friesians, 20 cows and 20 heifers, were used to study the effect of crude protein (CP) content of the whole ration on milk yield, milk composition and live-weight change when maize silage was fed as the basal ration. There were four treatments with five cows and five heifers on each. During lactation weeks 4–12 the cows on treatments 1, 2, 3 and 4 all received 7 kg maize silage dry matter (D.m.) plus 8 kg D.m./day of a concentrate containing either 14, 18, 22 or 24% CP, respectively; heifers received 1 kg/day less of both silage and concentrate. This produced whole-ration CP contents of 11·7, 13·9, 16.·0 and 17·1%. In the subsequent lactation weeks 13–20 silage feeding was increased to 9 kg D.m./cow/day and concentrate feeding decreased to 5 kg D.m./cow/day. Heifers again received 1 kg/day less of both silage and concentrate. This decreased whole-ration CP contents to 10·5, 11·7, 12·9 and 13·4%.

During lactation weeks 4–12 and 13–20, with the exception of milk fat content in weeks 13–20, there were significant linear effects of whole-ration CP content on milk yield, milk fat, protein, lactose and total solids. There were no significant curvilinear relationships. Thus, despite the fact that the highest numerical values were generally recorded for the animals on treatment 3, the results indicate that a whole-ration CP content of at least 17·1 and 13·4% are required in early and mid-lactation respectively.

In Expt 2 the loss of D.m., acid-detergent fibre and nitrogen from maize silage suspended in nylon bags in the rumen was measured. Compared with feeding either a low or high protein supplement, losses were greater for silage fed alone. If it is accepted that nitrogen loss can be approximated to protein degradability, then the value for maize silage was between 0·6 and 0·7. Using the same technique in Expt 3, comparable nitrogen losses for fish meal, decorticated groundnut meal and soya-bean meal were 0·3, 0·9 and 0·9, respectively, after 24 h incubation.

The relationship between calcium and phosphorus metabolism in wether sheep given high or low Ca diets, with or without 1α-hydroxycholecalciferol (1α-0H-D3) has been studied by a mineral balance and radioactive technique.

Ca absorption was not related to Ca intake but was stimulated by 1α-OH-D3. More Ca was absorbed by treated animals from the high Ca diet than from the low diet and all the extra Ca absorbed was retained, increased retention being brought about largely by an increase in the rate of bone accretion.

P absorption was increased to approximately the same extent from both diets suggesting that stimulation was due to the 1 α-0H-D3 treatment rather than increased Ca absorption. Whereas the extra P absorbed from the high Ca diet was retained, together with Ca, in bone and soft tissues, that absorbed from the low Ca diet was largely excreted in the urine. It is suggested that this difference in P retention reflects a difference in availability of Ca for retention in bone and P retention was in fact found to be directly related to Ca retention.

The roles of secretion of P into the gut, absorption of P from the gut and urinary excretion of P are discussed in relation to P homoeostasis.

As absorption of P from the intestine and loss of P to bone, soft tissues and urine increased, so endogenous faecal loss decreased until it reached a value of approximately 35 mg/day per kg body weight when it remained constant. It is suggested that this value may represent the inevitable loss of endogenous P in the faeces from a hay and concentrates diet and that this minimum value may have a bearing on the calculation of P requirements.

Results are reported from four separate trials carried out to determine the precision of the Sonatest (simple A-mode ultrasonic machine) and the Scanogram (modified linear scanner) for predicting the body composition of live cattle. Cattle in the four trials differed in breed, sex and origin, and the data provided an opportunity to determine the consistency of results in different circumstances. A total of 210 cattle were involved.

Fat thickness measurements (Sonatest and Scanogram) and fat and M. longissimus areas (Scanogram only) were taken at the 10th and 13th ribs and at the position of the 3rd lumbar vertebra. Their precision as predictors of carcass tissues percentages was examined when they were used in addition to live weight at evaluation.

There was little consistency between trials in the positions and measurements which gave the most precise prediction. The lowest within-breed residual standard deviations of carcass lean percentage obtained with fat thickness measurements taken by Sonatest were in the range 2·5–2·7 and there was little advantage in using additional measurements in multiple regression.

Fat areas taken by Scanogram were more precise predictors (within-breed residual standard deviations were close to 2·0). Precision was improved marginally to about 1·8 by using combinations of fat areas but M. longissimus areas were of little additional value.

Sucking behaviour and age of weaning were investigated in a herd of East African zebu cattle in which all calves born had been retained for several years. Suckling rates were compared at different periods of the 24 h cycle and found to be less at night than during any day-time period. The highest rates of suckling took place in the early morning and in the late afternoon. The early morning suckling was triggered by the dawn. During both night and day, suckling was irregular and bouts were presumably due to sympathetic induction because a high proportion of the calves was suckled on each occasion. The total sucking time and the number of bouts varied significantly in inverse relation to age of the calf but the duration of a sucking bout was constant at 8 min and independent of age. There was an average of 4·8 sucking bouts within 24 h giving a total duration of 38 min. Sucking behaviour did not exhibit sex-specific differences. The average age of natural weaning was 10 months but there was a marked sex difference. Female calves were prevented by their mothers from sucking after an average age of 8·8 months but male calves continued to be suckled for a further 2·5 months and were weaned at an average age of 11·3 months.

In a study of two breed groups, coarse wool Barki and ⅜ Merino ⅜ Barki, data were obtained on skin histology from birth to the age of 1 year, birthcoat fibre type arrays and fibre-follicle relationship.

The ⅜ Merino exceeded Barki lambs in the maximum S/P ratio, values of which were 3·40 and 5·64 obtained at 6 and 8 months in Barki and ⅜ Merino respectively.

Internal diameters of primary follicles ranged from 39·0 to 64·7 and from 41·0 to 56·8 μm; those of secondaries ranged from 19·8 to 34·9 and from 22·6 to 33·3 μm in Barki and ⅜ Merino respectively.

The ratio of primary to secondary follicle diameters showed high values in both breed groups; averages were 1·89 and 1·71 for internal diameter and 1·86 and 1·71 for external diameter in Barki and ⅜ Merino respectively.

Birthcoat fibre type arrays were mostly saddle, only 8·3% were plateau (P3) in both breed groups.

In Barki primary centrals grew halo hairs, super sickle (A, A′, B), primary laterals produced super sickle (A, A′, B), sickles, coarse and medium curly tips, and secondaries mostly grew medium and fine curly tips and histerotrichs.

In ⅜ Merino primary centrals produced halo hairs, super sickle (A, A′, B), primary laterals grew super sickle (A, A′, B), sickles, medium curly tips, and secondaries produced medium and fine curly tips and histerotrichs.

The self-fertile cultivar of swede rape (Brassica napus) Maris Haplona does not usually autopollinate. The entomophilie pollen grains cannot be transferred by wind alone; anthers when flicked by insects or artificially under dry conditions behave like catapults raising a cloud of pollen grains.

Variability in the concentration of sugar in its two types of nectaries (within a single flower) may increase the range of visiting insects. All visiting insects, including honey bees, carried out pollination; it is suggested that beehives should be placed in fields of oil-seed rape in order to increase pollination.

Field beans (Vicia faba L.) were given three irrigation treatments in the field during two seasons to examine reduction in leaf area by water shortage, through effects on number of cells, cell volume and cell arrangement.

The two crops behaved differently in the two seasons, presumably because of differences in environmental conditions. Season 1976 was brighter, warmer and drier than 1977. Water shortage reduced leaf area, dry matter and economic yield. The rate of leaf appearance and leaf longevity were also decreased by water shortage. The greatest biological and economic yields were produced by the wet treatment, followed by the medium and dry treatments. The treatment open to natural rainfall yielded least in 1976, but the second most in 1977. The yields of 1976 were considerably less than those of 1977.

The mineral content of N, P and K in leaves, stems, roots and seed of field beans, subjected to four watering regimes were compared in 1976 and 1977, and of a foliar nutrient spray in the latter year.

N, P and K content increased in all treatments with the increase in plant size and then declined at about the time of podding and death of the leaves.

Water shortage reduced the amounts of N, P and K throughout the growth period and for the seed at final harvest; the decrease was more pronounced in 1976 than in 1977. The plants under stress conditions benefited from the foliar nutrient spray more than the unstressed plants.

Four mature castrate Malawi goats, mean live weight 29·66 kg, were offered diets of poor quality Rhodes grass hay, or hay plus a concentrate comprising either maize bran or maize bran-cotton seed cake mixed in varying proportions, so that N intake varied from 4·76 to 20·81 g/day in a series of four metabolism trials.

By extrapolation metabolic faecal N loss was estimated to be 9·94 g/kg faecal D.M. and that of endogenous urinary N to be 0·121 g/kg W0·75. The daily maintenance requirement for nitrogen was shown by a factorial method to be 0·682 gN/kg W0·75, and by use of regression equations to be 0·688 g N/kg W0·75 or 0·314 g apparent digestible N/kg W0·75. The mean daily requirement was 8·8 g N, with a mean apparent digestibility of N of 43·4%, and a true digestibility of 80·0%.

A study was made to determine the effects of feeding, arginine injection (0·5 g/kg live weight) and insulin injection (0·5 i.u./kg live weight) on plasma insulin, prolactin and glucose concentrations in ten growing castrated male lambs. The effect of feeding on thyroxine concentrations was also determined. Each test was performed once in each of four periods which commenced at 84, 98, 113 and 127 days of age respectively.

Feeding stimulated the release of insulin and this response was larger during period 4 than period 1. The release of prolactin following feeding did not change with age. Thyroxine concentrations, which were generally higher in periods 1 and 2 than 3 or 4, decreased from 1 to 12 h after feeding.

Arginine injection stimulated a similar insulin response in lambs of all ages but the prolactin release following arginine injection diminished with age.

Insulin injection caused marked hypoglycaemia in all periods, followed by dramatic increase in prolactin concentrations which increased with age.

Over three grazing seasons a sward of perennial ryegrass, cv. Gremie, was grazed on a rotational paddock system by growing beef cattle weighing on average 185·5 kg when turned out to grass in April. Six grazing treatments were compared: grazing at 20- or 35-day intervals at each of three herbage allowances of 5·3, 4·3 and 3·3 kg grass D.M./100 kg live weight (4·0, 3·2 and 2·4 kg D.M./IOO kg live weight in the 1st year).

Measurements of net herbage accumulation and herbage consumption throughout each grazing season were made by taking grass sampling clips to ground level before and after grazing. Animal live-weight gain was assessed by regression on weighings of stock taken either every 7 or every 4 days depending on regrowth interval treatment. Adjustment of allocation treatment was made by varying the weight of stock grazing on a fixed paddock area.

In no year did annual net herbage organic-matter accumulation (NHA) differ significantly according to length of grazing interval. NHA showed a 4·7 % advantage for the longer interval in the first 2 years which changed to a deficit of 4·1 % in the final year of the experiment. Herbage allowance treatment produced no significant effect on NHA in any year. The interaction effect of herbage allowance and grazing interval was not significant.

Herbage consumption was greater in 2 years out of three under the 20-day regrowth period than with the 35-day treatment, although average degree of defoliation was slightly greater (44·0 v. 38·4%) with the longer interval.

Animal live-weight gain was significantly (P < 0·05) affected by herbage allowance in all 3 years; decreasing herbage allowance generally resulted in reduced live-weight gain. Only in 1 year was there a significant difference (P < 0·001) in live-weight gain with the different rotation length treatments.

In three field trials, three varieties of cow pea (Prima, Ife Brown and Vita-5) grown as mono-crops or interplanted with maize (in one trial) were sprayed with a combination of PKS or NPKS in solutions containing 100, 100, 100 and 41 mg/1 of N, P, K and S respectively at the onset of flowering or 2 weeks later, in order to evaluate the effect of foliar spray of mineral nutrients on the performance of the crop.

Early season sprayings of NPKS and PKS increased vegetative development and dry-matter accumulation in both mono- and mixed-croppings of Prima and Vita-5, but had little effect on Ife Brown. Seed yield was increased by 80–225 kg/ha in Prima and 160–218 kg/ha in Vita-5. In all cases of enhanced growth and yield, spraying at flowering was more beneficial than spraying 2 weeks later whilst NPKS evoked greater responses than PKS.

In two late-season trials, all treatment combinations enhanced vegetative development and dry-matter accumulation in Ife Brown and Vita-5. Seed yield of Ife Brown was increased by 206 kg/ha in the first late season and by 550 kg/ha in the second late season whilst that of Vita-5 was increased by 226 kg/ha in the first late season, during which it was planted. In the first late season of Ife Brown, late spraying was more beneficial than early spraying, whilst plants sprayed with PKS out-yielded those sprayed with NPKS. On the other hand, in the second late season, Ife Brown responded more to early spraying than late spraying and NPKS was more beneficial than PKS: Vita-5 grown in the first late season responded similarly.

In all cases of enhanced seed yield during the seasons, the pathway of response was through increased pod production and/or retention on the plants. The results therefore suggest that foliar spray of nutrients during the post-flowering period enhance seed yield in cow pea by ensuring prompt delivery of mineral nutrients to the site of photosynthesis at a time when nutrient depletion in the leaves and reduced efficiency of mineral nutrient uptake by the roots limit the supply of photosynthates to developing reproductive organs.

Sowing winter oil-seed rape cv. Victor late in autumn (late September or October) in seven seasons from 1970 to 1977 gave enormously variable seed yields, from 120 to 450 g/m2. All crops made little growth before winter, and yield was related to the size of the crop at flowering, a function of the length of time for radiation interception and growth between the ‘beginning of spring’, when mean temperatures rose consistently above 5 °C, and full flower in late May. A late spring in 1970 gave the poorest growth and lowest yield, whereas in 1977 an early spring coincided with late flowering to give exceptional growth, and yields higher than from any early sowing.

Crops sown in early autumn (before mid-September) produced more consistent seed yields, 280–360 g/m2, except in the dry year of 1976. All grew well in autumn, overwintered with a large leaf area, and once temperatures rose in spring, rapidly reached peak area and full flower in early May. They were all large at flowering, and yield was apparently limited more by post-flowering events.

With all sowings numbers of pods and seeds were largely determined during a 3-week phase in late May and early June, extending from full flower until most pod hulls had finished growing. Late sowings produced 3000–6000 pods/m2, and the number of seeds retained per pod varied widely, from 7 on a poorly grown crop to 22 on a well grown crop, thus expressing the yield potential determined by crop size at flowering. Early sowings, however, produced apparently excessive numbers of pods (6000–12000/m2) and few seeds per pod (6–10), so that yield varied little, regardless of crop size. Early in the phase, when the number of seeds was determined, the mass of yellow flowers at the top of the crop reflected or absorbed up to 60 % of incoming radiation, and then the large number of pods increasingly shaded each other and competed for assimilate, resulting in heavy seed losses. A high-yielding crop type may therefore need to incorporate the restricted pod production and good seed retention of some well-grown latesown crops with the reliability and desirable agronomic features of early-sown crops.

Final seed weight varied more between seasons (3·7–5·3 nag) than between sowings. Seed growth mainly took place after the number of seeds had been determined, the duration depending on temperature, but rate of growth apparently more on assimilate supply, a function of environmental factors and the number of competing seeds.

The aim of the experiments reported here was to investigate the possibility of minimizing yield losses due to late sowing by increasing leaf area and radiation interception during the pre-flowering period. The use of large seed did increase leaf area, radiation interception and crop weight at flowering. Normally this would be expected to lead to increased yield of late sowings, but hot, dry conditions severely affected all crops during seed growth. Applying fertilizer nitrogen in the seed bed boosted pre-flowering growth of early but not late sowings, probably because low temperatures prevented a response in the latter. Increasing plant population density did improve leaf cover and radiation interception until just before flowering, but leaf canopies in late-sown, low density crops were then able to expand more and function for longer owing to less shading by flowers and pods. Seed retention was improved and, even with as few as 8 plants/m3 in 1973–4, a worthwhile yield was still obtained.

Four experiments were conducted to determine whether the products of fermentation in maize silage could limit food intake by lambs. In all experiments the lambs were fed ad libitum adietoi dried and pelleted lucerne, but were deprived of food for 4 h on the morning of intraruminal infusion. The infusions were made according to latin-square designs and commenced after 1 h of food deprivation. In Expt 1, the infusion of 3·5 I acetic acid solutions to provide 15, 30 and 60 g acetic acid had no effect (P > 0·05) on cumulative food intake compared with water. In Expt 2 infusion of three maize silage extracts differing in their composition of nitrogenous constituents suppressed (P < 0·05) food intake relative to acetic acid alone, but there were no significant (P < 0·05) differences among the extracts. When 2·71 of an extract from a low dry matter (D.M.) maize silage were infused in Expt 3 cumulative food intake was significantly (P > 0·05) decreased compared with acetic acid, but ruminal osmolality was increased to 430 mosmol/kg and ruminal pH declined to less than 5·5. In Expt 4, 1·5 1 of the low D.M. maize silage extract, an amount actually consumed by lambs during 4 h, decreased voluntary intake by 25% compared with the control but the effect was not statistically significant. Estimates of ruminal osmolality and pH ranged between 250 and 360 mosmol/kg and between 68 and 6·2 respectively. Measures of acid-base status were not affected (P < 0·05) by the infusion of acetic acid or maize silage extract. Whether specific soluble constituents of maize silage fermentation could limit food intake by ruminants was not proven.

Five rumen-cannulated Cheviot lambs were used in two successive experiments to differentiate the effect of osmolality from the specific effects of the soluble compounds in maize silage on voluntary intake and ruminal motility. In Expt 1, extracts from fresh (FCE) and ensiled (CSE) maize were adjusted to pH 6·5 andtonicities of 200, 400, 800, and 1600 m-osmol/kg with NaCl, then infused into the rumen according to a 5×5 repeated latin-square design. Water was infused as the control. In Expt 2 solutions of NaCl, with tonicities similar to those in Expt 1, were also infused into the rumen. The lambs were fed ad libitum a diet of pelleted lucerne, but on the morning of infusion they were deprived of food for 4 h and infused continuously for 4 h commencing after 1 h of food deprivation. The infusion of FCE and CSE resulted in no significant difference in cumulative dry matter (D.M.) intake. However, increasing the tonicities of the infusates caused a linear increase (P < 0·05) in ruminal osmolality and a corresponding linear reduction in food intake during the first 30 min of feeding. Infusion of NaCl alone produced a similar effect on intake. The linear regression coefficients of food intake (Y) (g D.M./kg W0·75) against rumen osmolality (X) (m-osmol/kg) were –0·077±0·0085 (r2 = 0·78), –0·082±0·0095 (r2 = 0·76) and –0·073±0·0065 (r2 = 0·84) for FCE, CSE and NaCl, respectively, with no significant differences among the infusates. The frequencies of ruminal contractions during the infusion of FCE and CSE were similar (P > 0·05) and were unaffected (P > 0·05) by ruminal osmolality up to 550 m-osmol.

A breeding programme was initiated to develop populations of sunflower (Helianthus annuus L.), suitable for the semi-arid regions of Iran. Crosses were made between Chernianka 66 (having low yield, low oil content, short stem, and early maturity), as the female parent and VNIMK 8931 and Record (having high yield, high oil content, long stem, and late maturity) as the male parents. Selection, together with selfing and sib-crossing, resulted in six populations with high yields, high oil content, short stems, and early maturity.

Correlation coefficients of oil yield and its components, i.e. seed yield and oil content, were highly significant and significant respectively. Capitulum diameter had highly significant correlations with seed yield and oil yield. Earlier maturity was associated with shorter stems and lower oil content.

Individual leaf blades of onion plants were treated with 14CO2 57, 71, 93 and 112 days after transplanting. The distribution of labelled translocates was determined 24 h after labelling. The older the plant the more of 14C-assimilate left the source leaf; 112 days after transplanting, leaves 1–7 all exported more than 70% with leaf 7 showing a maximum export of 93 % of the fixed carbon. In the early stages of the development, the leaves, pseudostem and the bulbs were the major sinks in decreasing order. But the bulbs became the dominant sink at 71 days and at 112 days the bulbs almost monopolized the import of the radiocarbon, with 90 % of the exported 14carbon being recovered from them. The bulbs received their 14carbon supply from all the leaves but particularly the basal and middle region leaves. Roots were a weak sink and received their 14carbon supply only from basal leaves. The stem and mature leaf blades were unimportant sinks throughout the life of the plant.