In this review, I discuss the possibility that dying cells produce much of the auxin in vascular plants. The natural auxin, indole-3-acetic acid (IAA), is derived from tryptophan by a two-step pathway via indole pyruvic acid. The first enzymes in the pathway, tryptophan aminotransferases, have a low affinity for tryptophan and break it down only when tryptophan levels rise far above normal intracellular concentrations. Such increases occur when tryptophan is released from proteins by hydrolytic enzymes as cells autolyse and die. Many sites of auxin production are in and around dying cells: in differentiating tracheary elements; in root cap cells; in nutritive tissues that break down in developing flowers and seeds; in senescent leaves; and in wounds. Living cells also produce auxin, such as those transformed genetically by the crown gall pathogen. IAA may first have served as an exogenous indicator of the presence of nutrient-rich decomposing organic matter, stimulating the production of rhizoids in bryophytes. As cell death was internalized in bryophytes and in vascular plants, IAA may have taken on a new role as an endogenous hormone.

1 Although much is known about the effects of plant hormones and their role in the control of growth and differentiation, little is known about the way in which hormone production is itself controlled or about the cellular sites of hormone synthesis. The literature on hormone production is discussed in this review in an attempt to shed some light on these problems.

2 The natural auxin of plants, indol-3yl-acetic acid (IAA) is produced by a wide variety of living organisms. In animals, fungi and bacteria it is formed as a minor by-product of tryptophan degradation. The pathways of its production involve either the transamination or the decarboxylation of tryptophan. The transaminase route is the more important.

3 In higher plants auxin is also produced as a minor breakdown product of tryptophan, largely via transamination. In some species decarboxylation may occur but is of minor important. Tryptophan can also be degraded by spontaneous reaction with oxidation products of certain phenols.

4 The unspecific nature of the enzymes involved in IAA production and the probable importance of spontaneous, non-enzymic reactions in the degradation of tryptophan make it unlikely that auxin production from tryptophan can be regulated with any precision at the enzymic level. The limiting factor fro auxin production is the availability of tryptophan, which in most cells is present in insufficient quantities for its degradation to occur to a significant extent. Tryptophan levels are, however, considerably elevated in cells in which net protein breakdown is taking place as a result of autolysis.

5 An indole compound, glucobrassicin, occurs in Brassica and a number of other genera. It breaks down readily to form a variety of products including indole acetonitrile, which can give rise to IAA. There is, however, no evidence to indicate that glucobrassicin is a precursor to auxin in vivo.

6 Conjugates of IAA, e.g. IAA-aspartic acid and IAA-glucose, are formed when IAA is supplied in unphysiologically high amounts to plant tissues. These and other IAA conjugates occur naturally in developing seeds and fruits. There is no persuasive evidence for the natural occurrence of IAA-protein complexes.

7 Tissues autolysing during prolonged extraction with ether produce IAA from tryptophan released by proteolysis. IAA is produced in considerable quantities by autolysing tissues in vitro.

8 During the senescence of leaves proteolysis results in elevated levels of tryptophan. Large amounts of auxin are produced by senescent leaves.

9 Coleoptile tips have a vicarious auxin economy related to IAA from the seed. These move acropetally in the xylem and accumulate at the coleoptile tip. The production of auxin in coleoptile tips involves the hydrolysis of IAA esters and the conversion of labile, as yet unidentified compounds, to IAA. There is no evidence for the de novo synthesis of IAA in coleoptiles.

10 Practically all the other sites of auxin production are sites of both meristematic activity and cell death. The production of auxin in developing anthers and fertilized ovaries takes place in the regressing nutritive tissues (tapetum, nucellus, endosperm) as the cells break down. In shoot tips, developing leaves, secondarily thickening stems, roots and developing fruits auxin is produced as a consequence of vascular differentiation; the differentiation of xylem cells and most fibres involves a complete autolysis of the cell contents; the differentiation of sieve tubes involves a partial autolysis. There is no evidence that meristematic cells produce auxin.

11 The lysis ad digestion of cells infected with fungi and bacteria results in elevated tryptophan levels and the production of auxin. Viral infections reduce the levels of tryptophan and are associated with reduced levels of auxin.

12 Crown-gall tissues produce auxin. It is suggested that the crown-gall disease may involve at any given time the death of a minority of the cells which produce auxin and other hormones as they autolyse; the other cells grow and divide in response to the hormones.

13 Auxin is produced in soils, particularly those rich in decaying organic matter, by micro-organisms. This environmental auxin may be important for the growth of roots.

14 There is no convincing evidence that auxin is a hormone in non-vascular plants. The induction of rhizoids in liverworts by low concentrations of auxin can be explained as a response to environmental auxin.

15 Abscisic acid is synthesized from mevalonic acid in living cells. It is possible that under certain circumstances, abscisic acid or closely related compounds are formed by the oxidation of carotenoids.

16 The sites of gibberellin production are sites of cell death. It is possible that precursors of gibberellins, such as kaurene, are oxidized to gibberellins when cells die.

17 Cytokinins are present in transfer-RNA (tRNA) of animals, fungi, bacteria and higher plants. They are probably formed in plants by the hydrolysis of tRNA in autolysing cells. There is evidence that they are also formed in living cells in root tips.

18 Ethylene is produced in senescent, dying or damaged cells by the breakdown of methionine.

19 It was shown many years ago that wounded and damaged cells produced substances which stimulate cell division. It now seems likely that the production of wound hormones and the normal production of hormones as a consequence of cell death are two aspects of the same phenomenon. Wounded cells can produce auxin, gibberellins, cytokinins and ethylene.

20 The control of hormone production in living cells is a biochemical problem which remains unsolved. The control of production of hormones formed as a consequence of cell death depends on the control of cell death itself. Cell death is controlled by hormones which are themselves produced as a consequence of cell death.

21 In spite of the fact that dying cells are present in all vascular plants, in all wounded and infected tissues, in certain differentiating tissues in animals, in cancerous tumours and in developing animal embryos, the biochemistry of cell death is a subject which has been almost completely ignored. Dying cells are an important source of hormones in plants; some of the many substances released by dying cells may also be of physiological significance in animals.

A re-examination of the evidence for auxin production by coleoptile tips reveals that it is not conclusive and that several important problems remain unresolved. The possibility that auxin and auxin precursors move acropetally in the xylem was tested by analysing guttation fluid from intact coleoptiles, decapitated coleoptiles and primary leaves of Avena sativa. In all cases two zones of auxin activity were detected on chromatograms of the acidic ether-soluble fraction, one of which corresponded to the Rf of indol-3-yl acetic acid (IAA). Similar auxin activity was found in guttation fluid from seedlings of Zea mays, Triticum aestivum and Hordeum vulgare. Evidence that guttation fluid also contains alkali-labile auxin complexes was obtained. Experiments on the movement of dyes and radioactive IAA introduced into the xylem of transpiring or guttating coleoptiles showed that these substances accumulate at the tip of the coleoptile, or at the apical region of decapitated coleoptiles. The hypothesis that IAA and 'inactive' auxins move acropetally in the xylem from the seed to the coleoptile tip where they accumulate and where the 'inactive auxins' can be converted to IAA is shown to be consistent with the classical work on coleoptiles; it can also explain the autonomous curvature of coleoptiles and the influence of the roots on the auxin contect of coleoptile tips. An analogous accumulation of auxin probably occurs at the tips of primary leaves. The anomalous auxin economy of coleoptile tips is discussed.

Cambium and differentiating xylem and phloem tissues from the trunks of trees of Acer pseudoplatanus L., Fraxinus excelsior L., and Populus tremula L. were extracted with ether and tested for auxin, which was found on chromatograms of the acidic fraction at an Rf corresponding to that of indol-3yl-acetic acid in five solvent systems. In addition, small amounts of auxin with a higher Rf in ammoniacal isopropanol were found in phloem samples. The amounts of auxin were greatest in xylem samples, less in the cambium, and least in phloem. The differences, which cannot be explained in terms of differential losses during extraction and purification, suggest that auxin is actually formed in differentiating xylem tissue. The significance of these results is discussed.

Auxin was detected in samples of substrata supporting bryophytes in a variety of locations in both Britain and Malaya. Activity occurred on chromatograms at zones corresponding to the Rf of indole acetic acid. The range of concentrations found, 0.4-10.4ug/1, probably represents a two-to five-fold underestimate due to losses during extraction and purification. The amounts of auxin in samples of soil on which bryophytes were not growing were within the same range. The importance of this environmental auxin for the induction of rhizoids in liverworts and for roots of higher plants is discussed.

Autolysing plant tissues are known to produce auxin when extracted with ether. It has been shown that autolysing plant, yeast and rat liver tissues produce auxin in vitro; this suggests that relatively unspecific mechanisms are involved. Furthermore, sterile plant and animal tissues which have been killed by freezing and thawing induce nodules of differentiated cells in a previously undifferentiated callus of Phaseolus vulgaris. The callus tissue is known to differentiate in response to applied gradients of auxin. Plant and animal tissues killed by boiling were considerably less effective in inducing differentiation in the tissue. The evidence indicates that auxin is a normal product of autolysing cells. It is suggested that dying cells are an important source of auxin in the plant.

In senescent leaves proteins are hydrolysed to amino-acids and peptides, which might be expected to release protein-bound auxin and also to provide considerable amounts of trypotophan which can be converted by many plant tissues to the auxin indolyl-3-acetic acid (IAA). We have therefore investigated the concentrations of auxin in senescent leaves.

Mature trifoliate leaves from plant of Phaseolus vulgaris and leaves from young plants (2-3 weeks old) of Avena sativa were detached and placed with their petioles or bases in distilled water in the dark at 25° C. In these conditions, the leaves become senescent and turn yellow. Samples were taken at various times (at intervals of 1 or 2 days), weighed and stored in the deep freeze until they were extracted with peroxide-free ether for 3 h at 0° C. The ether extract was partitioned and the acidic fraction was run on paper chromatograms with isopropanol : ammonia : water (8:2:1 v/v). The zone corresponding to IAA was eluted and the auxin was estimated using an Avena coleoptile straight growth bioassay. The amounts of auxin extracted from the leaves at various times are shown in Figs. 1 and 2.

It can be seen that in both cases there is a large increase in the amount of auxin present over a period of 6 days. The amounts measured represent the resultant of auxin production and auxin destruction: in the case of Avena, after about the fourth day the rate of destruction exceeds the rate of production. The fall in total auxin was observed in each of six experiments.

The level of auxin in leaves and petioles is involved in the control of abscission so the production of auxin by senescent leaves, if it is a general phenomenon, may be an important factor which so far has been overlooked.

The formation of callus at the basal end of tobacco internode tissues cultured on a basic medium has been used as an indication of the presence of auxin within the tissues. It has been shown in this way that sections of internode are capable of producing auxin. This production of auxin is related to the continued activity of the vascular cambium. If cambial activity and vascular differentiation are eliminated, auxin is no longer produced. When tissues in which cambial activity and vascular differentiation are taking place are cultured on a medium containing an inhibitor of polar auxin transport, tri-iodo benzoic acid, serried ranks of xylem tracheids are formed. It is suggested that auxin is produced as a consequence of xylem differentiation and the observations reported in this paper are interpreted in the light of this hypothesis. It is also suggested that kinins may be produced as a result of xylem and phloem differentiation, and the possibility that autolysing cells are a major source of both auxins and kinins in the plant is discussed.

Segments of mesocotyls of Avena sativa L. transported (1-14C) indol-3yl-acetic acid (IAA)with strictly basipetal polarity. Treatment of the segments with solutions of sorbitol caused a striking increase in basipetal auxin transport, which was greatest at concentrations around 0.5M. Similar effects were observed with mannitol or quebrachitol as osmotica, but with glucose or sucrose the increases were smaller. Polar transport was still detectable in segments treated with 1.2M sorbitol. The effects of osmotic stress on the polar transport of auxin were reversible, but treatment with sorbital solutions more concentrated than 0.5M reduced the subsequent ability of mesocotyl segments to grow in response to IAA. The increased transport of auxin in the osmotically stressed segments could not be explained in terms of an increased uptake from donor blocks. The velocity of transport declined with higher concentrations of osmoticum. The reasons for the enhancement of auxin transport by osmotic stress are not known.

Auxin (IAA) transport was investigated using crown gall suspension tissue culture cells. We have shown that auxin can cross the plasmalemma both by transport of IAA anions on a saturable carrier and by passive (not carrier-mediated) diffusion of the lipid-soluble undissociated IAA molecules (pK=4.7). The pH optimum of the carrier for auxin influx is about pH6 and it is half-saturated by auxin concentrations in the region of a 1-5u-M. We found that the synthetic auxin, 2,4D specifically inhibited carrier-mediated IAA anion influx, and possibly also efflux. Other lipid-soluble weak acids which are not auxins, such as 3,4-dichlorobenzoic acid, had no effect on auxin transport. By contrast, we found that TIBA, an inhibitor of polar auxin transport in intact tissue inhibited only the carrier-mediated efflux of IAA.

When the pH outside the cells is maintained below that of the cytoplasm (pH7), auxin can be accumulated by the cells: In the initial phase of uptake, the direction of the auxin concentration gradient allows both passive carrier-mediated anion influx (inhibited by 2,4D) and a passive diffusion of undissociated acid molecules into the cells. Once inside the cytoplasm, the undissociated molecules ionise, producing IAA anions, to a greater extent than in the more acidic extra-cellular environment. Uptake by passive diffusion continues as long as the extra-cellular concentration of undissociated acid remains higher than its intra-cellular concentration. Thus, the direction of the auxin anion concentration gradient is reversed after a short period of uptake and auxin accumulates within the cells. The carrier is now able to mediate passive IAA anion efflux (inhibited by TIBA) down this concentration gradient even though net uptake still proceeds because the carrier is saturable whereas passive diffusion is not.

Auxin 'secretion' from cells is regarded as a critical step in polar auxin transport. The evidence which we present is consistent with the view that auxin 'secretion' depends on a passive carrier-mediated efflux of auxin anions which accumulate within the cells when the extra-cellular pH is below that of the cytoplasm. The implications of this view for theories of polar auxin transport are discussed.

The original, basipetal polarity of auxin transport persisted in the stems of inverted cuttings of Tagetes, tomato and tobacco in spite of the reversal of the relative positions of the roots and shoots. No significant acropetal auxin transport could be detected even after four months growth. These results indicate that the polarity of newly formed cells in secondarily thickening internodes is determined by the existing polarity of auxin transport within the tissues.

Auxin transport was investigated in excised stem segments of Nicotiana tabacum L. by the agar block technique using (I-14C) indol-3yl-acetic acid (IAA). The ability of the stems to transport auxin basipetally increased as secondary development proceeded; by contrast the ability of the pith to transport auxin declined with age. By separation of the stem tissues it was shown that the great majority of auxin transport took place in cells associated with the internal phloem and in cells close to the cambium; in both cases similar velocities of transport were found (c 5.0 mm h-1 at 22°C). The effects of osmotic gradients on auxin transport through the internal phloem were investigated. IAA was found by chromatography to account for practically all the radioactivity in receiver blocks and ether extracts of stem segments. The significance of these results is discussed.

The uptake of the auxin indol-3-yl acetic acid (IAA) into plant cells is of interest not only because this compound is a hormone, but also because its movement across the plasma membrane is probably involved in the polar transport of auxin. The plasma membrane contains auxin binding sites and may be a primary site of hormone action.

IAA partitions into non-polar solvents from acidified aqueous solutions because the undissociated acid is more soluble in such lipid solvents than in water. There is known to be a passive, non-metabolic component of the uptake of IAA and of the synthetic auxin 2,4-dichlorophenoxyacetic acid (2,4-D) into plant tissue which has been ascribed to the diffusion of the undissociated acid across the plama membrane. A carrier-mediated mechanism for auxin anion uptake is also possible but has not been conclusively demonstrated.

Uptake by the diffusion mechanism is linearly related to the concentration of the undissociated acid which is a function of the acid's pK and the pH of the incubation medium. If the pH of the medium is lower than that of the cells, the cells accumulate weak acid; the equation requires that the concentration of undissociated acid should be the same in each compartment. Thus the relation between the initial rate of uptake and pH should resemble a dissociation curve with a midpoint at the pK of the weak acid. This prediction is realized for the uptake of benzoic acid (pK=4.2) by yeast but not by the bacterium Proteus vularis, when, although the curve is still that of a dissociation, its midpoint is displaced by 1 pH unit above the pK of benzoic acid. Such displacement seems fairly widespread. By collating data from ninety experiments on pH dependence of biological effects of weak acids, a composite curve is obtained relating pH to log concentration of acid required to give a standard response; the midpoint of the curve is at a higher pH than the pK. Data on IAA and 2,4-D uptake reveal a similar effect. Here we suggest an explanation of this displacement which may be of general biological significance.

Almost nothing is known about the establishment of cellular polarity underlying the polar auxin transport system of higher plants. Osborne has suggested that the apical ends of cells derived from an apical meristem by sequential divisions are younger than the basal ends: their polarity and the basipetal transport of auxin are due to this age difference. Sachs in his work on regenerating vascular strands has found that gradients of auxin may be responsible for establishing the cellular polarity and the subsequent transport of auxin in the direction of the initial gradient. Shoot tips and expanding dicot leaves contain relatively high levels of auxin. The basipetal polarity of auxin transport in petioles and stems is therefore associated with basipetal auxin gradients. In grass coleoptiles the greatest amounts of auxin are found at the tip, where basipetal auxin transport is also associated with basipetal auxin gradients."

In monocot leaves which grow by a basal intercalary meristem, the pattern of cell division and of auxin distribution is more or less the reverse of that found in shoot tips. Sequential divisions of the basal meristem presumably make younger the basal ends of cells; and in growing monocot leaves the greatest amounts of auxin are found at the base. The polarity of auxin transport in monocot leaves is therefore of considerable interest.

Hertel and Leopold reported that in the primary leaf of Zea mais, auxin transport was basipetal. No other references to auxin transport in monocot leaves are available and I therefore tested the leaves of a number of species. In every case auxin transport was basipetal.

In leaves of young plants of Avena sativa, basipetal auxin transport took place across the meristematic region at the base of the leaf and also in the leaf sheath, which grows by a basal meristem. Plants germinated and grown in darkness yielded similar results. Less auxin transport was found near the leaf tip than in the younger, more basal parts of the leaf and younger leaves had a greater ability to transport auxin than older leaves. A decline in the ability of cells to transport auxin as they grow older has been observed in a number of other species and tissues.

Short-duration pigeonpea can give up to three harvests in environments with mild winters (eg. minimum temperature above 10*C) such as those prevailing in peninsular India (Sharma, Saxena & Green, 1978; Chauhan, Venkataratnam & Sheldrake, 1984). This is mainly due to the short time (about 120 days) taken to produce the first flush, and the strong perennial character of pigeonpea. The seed yield of short-duration pigeonpea in this multiple-harvest system may reach 5.2t/ha (Chauhan et al. 1984).

Venkataratnam & Sheldrake (1985) found that the yield of the second harvest of medium-duration pigeonpea was significantly influenced by the method of harvesting of the first flush. The lower the plants were cut, the smaller were the second-harvest yields. A positive relationship between the height at which the stem was cut and success of ratooning was also reported by Suarez & Herreara (1971). Tayo (1985), however, found that in the lowland tropics, plants of a dwarf pigeonpea variety ratooned at 0.3 m had better growth and yield than hand-picked plants; ratooning at 0.6 m height was intermediate. Information on the effect of different harvest methods on yield of short-duration pigeonpea in subtropical, semi-arid environments is not available. The objective of this study was to obtain this information.

Environmental and cultural factors that may limit the yield of short-duration pigeonpea were investigated over three seasons. Plants in the peninsular Indian environment at Patancheru grew less and produced less dry matter by first-flush maturity than at Hisar, a location in northern India where the environment is considered favourable for the growth of short-duration pigeonpea. However, with a similar sowing date in June, the mean seed yields of three genotypes, ICPL4, ICPL81 and ICPL87, were very similar, at about 2-3t/ha, in both environments. This was mainly due to the higher ratio of grain to above-ground dry matter at Patancheru. In addition to the first harvest, all genotypes showed a potential for two more harvests owing to the warm winters at Patancheru. The potential for multiple harvests was particularly high in ICPL 87, which yielded 5.2t/ha from three harvests in 1982-3, 3.6t/ha from two harvests in 1983-4, and 4.1 t/ha from three harvests in 1984-5. The optimum plant population density at Patancheru was 25-35 plants/m2 for ICPL 87, but was higher for the other two genotypes.

At Patancheru, the total dry-matter and seed yield of first and subsequent harvests were significantly reduced by delaying sowing beyond June. Generally, the second-and the third-harvest yields were lower on vertisol than on alfisol under both irrigated and unirrigated conditions.

The total yield of ICPL 87 from two harvests was far higher than that of a well-adapted medium-duration genotype BDN 1, grown over a similar period. The yield advantage was greater on the alfisol because of the better multiple harvest potential of this soil. The results of this study demonstrate that properly managed short-duration genotypes of pigeonpea may have considerable potential for increased yield from multiple harvests in environments where winters are warm enough to merit continued growth.

The feasibility of growing pigeonpea [Cajanus cajan (LInn.) Millsp.] as a perennial crop was investigated during 1980-82. The medium-duration pigeonpea genotype 'ICP 1-6', sown in the post-rainy season at a population of 30 plants/m2, was allowed to perennate for 18 months, during which it produced 3 flushes of pods at 5,15 and 18 months after sowing. There was a substantial plant mortality after the first-flush harvest, but due to the high-sowing rate many plants survived and regenerated to form a closed canopy in the following rainy season. The mean yield of 2 seasons was 0.5 tonne/ha in the first flush, 1 tonne/ha in the second and 0.05 tonne/ha in the third. The yield from the second flush was obtained without weeding or insecticide spray and was comparable to that of the rainfed crop of medium-duration genotypes. Considerable leaf fall also occurred, which contributed 40kg N/ha to the soil. The yield from the third flush was very low to warrant continuation of the crop for another 3-4 months after the second-flush crop. At this harvest the mean firewood (air-dried stem) yield was 3.5 tonnes/ha. The system has good potential in the wet rainy season fallows in peninsular India, as it enables pigeonpea after the rainy season with little efforts and few inputs.

In Peninsular India medium duration pigeonpeas (Cajanus cajan) are normally sown soon after the onset of the monsoon, in June or July; they mature around December, when they are usually cut down and removed from the field. However, if they are harvested by ratooning or by picking the pods, the plants go on to produce a second flush of pods, which matures around March. In experiments conducted in four growing seasons at ICRISAT Center, second harvest yields were usually greater for non-ratooned than ratooned plants, and in experiments conducted on Vertisols they were greater for the plants ratooned high up in the plant than for those cut closer to the ground. Second harvest yields of non-ratooned plants without irrigation on Alfisols were on average 66% of the first harvest yields, but on Vertisols only 37%, in spite of the greater water-holding capacity of the latter. On Alfisols second harvest yields were approximately doubled by a single irrigation, but there was less response to irrigation on Vertisols. The poorer second harvest yields on Vertisols may have been due to the damaging effects of soil cracking on the root system of the plants. In non-ratooned plants from which the first and second flushes of pods were harvested together, yields were less than the total yield obtained from non-ratooned plants in two harvests, even though the yield loss, mainly due to pod shattering, was as little as 4% in one year. The taking of second harvests from pigeonpeas grown on Alfisols may have considerable potential as a method of obtaining additional yield for little extra cost.

During the 3 years 1974-77 we studied the anatomy of most of the tissues and organs of the pigeonpea and, in the course of this work, have built up a collection of permanent microscope slides. These are retained in the Anatomy Laboratory at ICRISAT as a reference collection and may be consulted by anyone who is interested.

This report contains a brief and preliminary description of pigeonpea anatomy. We have studied the anatomy of several different cultivars; unless otherwise indicated, the following general descriptions apply to all cultivars investigated. We have not noticed any striking qualitative anatomical differences among cultivars; nodoubt quantitative differences exist, but these are difficult to establish with anatomical methods involving very small samples.

Many of the features of the anatomy of the pigeonpea are similar to those of other dicotyledonous plants, described in standard textbooks of anatomy. We have not attempted to duplicate these descriptions. Some aspects of the anatomy of the pigeonpea have been covered in detail by Dr. P. Venkateshwara Rao in his Ph.D. thesis under reference (nodate). A copy of this thesis is available in the ICRISAT library.

Larger seeds of chickpea (Cicer arietinum) and pigeonpea (Cajanus cajan) gave rise to larger seedlings than did smaller seeds. When approximately half the cotyledonary reserves from pigeonpea seeds were removed, seedling weight was reduced to about half of the controls, suggesting that seedling growth was related to the reserve material in the seeds. Seed-grading had no significant effect on the yield of either of these crops grown on a Vertisol and on Alfisol in Andhra Pradesh, or on an Entisol in Haryana or in the Lahaul valley of the western Himalayas. Seeds harvested from pigeonpea grown from larger seeds were significantly heavier than those from plants derived from small seeds, probably because of the genetic heterogeneity of the varieties.

The influence of seed size on seedling growth of pigeonpea [Cajanus cajan (Linn.) Millsp.] and chickpea (Cicer arietinum Linn.) was investigated to predict probable consequnces of selection for seed size in breeding programmes. Seeds of 20 pigeionpea varieties with 100-seed weights of 4.5 to 22 g and 23 chickpea varieties with 100-seed weights of 5 to 32 g were sown in the field, and the leaf area and dry weight of the seedlings were measured at intervals up to 56 and 30 days respectively. In both species there was a close linear relationship between 100-seed weight and seedling weight (r = 0.77* for 14-day-old pideonpea; r = 0.82** for 16-day-old chickpea). In pigeonpea the relationship was even closer (r = 0.95**) when varieties having 100-seed weights of over 15 g were excluded. With the advancement of growth the closeness of these relationships declined. Large-seeded varieties of these crops produce larger and more vigorous seedlings, which will have an advantage in stand establishment under adverse conditions.

Pod photosynthesis is known to contribute to seed filling in a number of legume crops, and may also be of importance in chickpeas (Cicer arietium L.), which have green pods possessing stomata. Although the pods of chickpeas are borne in the leaf axils, they generally hang below the leaves and are consequently more or less shaded; but a few lines have recently been identified in which the pods are borne above the leaves. This *exposed pod* character could be incorporated into new cultivars by breeding if it were shown to be of advantage. The effect on yield and yield components of exposing pods of normal cultivars was investigated in field experiments at three locations in India: at Hyderabad and Hissar during the winter season, and in the Lahaul valley in the Himalayas during the summer season. A significant effect of pod exposure on yield or yield components was not observed in any of the experiments, except at Hissar where a slight but significant increase in 100-seed weight was noted. The *exposed pod* character is unlikely to be of use in breeding for higher yield potentials.

Genotypic differences exist in the sensitivity of cultivars of chickpea to iron deficiency. Sensitive cultivars exhibited typical iron deficiency symptoms when grown on calcareous soils with high pH. FeSO4 sprays (0.5%) corrected deficiency symptoms and increased yields by up to 50% in cultivars inefficient in iron utilization, but gave no increase in cultivars that were efficient.

Research conducted by ICRISAT at Hissar (representative of N. India) and Hyderabad (representative of peninsular India) on the growth, pod development, yield components, nutrient uptake, source/sink relationships, fertilizer and irrigation response, effects of intercropping, apex removal, row orientation, sowing pattern, plant density and seed size, cv. plasticity and cv. differences in germination, Fe chlorosis, salinity tolerance, heat tolerance and water stress response of chickpea is reviewed.

In chickpeas (Cicer arietinum) flowering and pod development proceed acropetally. In plants grown under normal field conditions at Hyderabad, in peninsular India, and at Hissar in north India, at successively apical nodes of the branches there was a decline in pod number per node, weight per pod, seed number per pod and/or weight per seed. The percentage of nitrogen in the seeds was the same in earlier- and later-formed pods at Hyderabad; at Hissar the later-formed seeds contained a higher percentage. Earlier- and later-formed flowers contained similar numbers of ovules. The decline in seed number and/or weight per seed in the later-formed pods of 28 out of 29 cultivars indicated that pod-filling was limited by the supply of assimilates or other nutrients. By contrast, in one exceptionally small-seeded cultivar there was no decline in the number or weight of seeds in later-formed pods, indicating that yield was limited by 'sink' size.

On branches of indeterminate cultivars of pigeonpea, flowering begins at the basal nodes and proceeds acropetally; in morphologically determinate cultivars, flowering begins on the apical racemes and proceeds basipetally. In cultivars of both types, within the racemes flowering proceeds acropetally. Under normal conditions more pods are set from earlier-formed flowers than from later-formed flowers, many of which are shed. Consequently the earlier-formed pods are found at the more basal nodes of racemes, and in indeterminate cultivars at the more basal nodes on the branches. The average weight of earlier- and later-formed pods, collected from the basal and apical nodes of the racemes or of the branches, was similar; so was the number of seeds per pod, the weight per seed and the nitrogen content of the seeds. This pattern differs from that found in most herbaceous legumes, where later-formed pods are smaller, and indicates that pigeonpeas set fewer pods than they are capable of filling. This behaviour may be related to the intrinsically perennial nature of pigeonpeas. The comparison of the weights of earlier-and later-formed pods could provide a simple screening procedure for identifying plants with an annual nature among existing cultivars or in breeders' lines.

In field experiments carried out at Hyderabad, India with early and medium-duration cultivars of Cajanus cajan sown at the normal time, in July, removal of all flowers and young pods for up to 5 wk had little or no effect on final yield. The flowering period of the deflowered plants was extended and their senescence delayed. The plants compensated for the loss of earlier-formed flowers by setting pods from later-formed flowers; there was relatively little effect of the deflowering treatments on the number of seeds per pod or weight per seed. The plants were also able to compensate for the repeated removal of all flowers and young pods from alternate nodes by setting more pods at the other nodes.

The removal of flowers from pigeonpeas grown as a winter crop resulted in yield reductions roughly proportional to the length of the deflowering period, probably because maturation of these plants was delayed and occurred under increasingly unfavourable conditions as the weather became hotter.

The growth and development of two early (Pusa ageti and T-21) and three medium- duration (ST-1, ICP-1 and HY-3C) cultivars of pigeonpea (Cajanus cajan) were compared at Hyderabad, India, in 1974 and 1975; in 1976 cv. ICP-1 was studied. The pigeonpeas were grown on a Vertisol and on an Alfisol. The crop growth rate in the first 2 months was low. The maximum rate of 171 kg/ha/day was found in the fourth month of growth of cv.ICP-1 on Alfisol. The early culitvars, one of which (cv. Pusa ageti) was morphologically determinate, and the other (cv. T-21) indeterminate, did not differ in the proportion of dry matter partitioned into seeds. The mean dry weight of the above- ground parts of the medium cultivars on Vertisol in 1975 was 8.45 t/ha, including 2.23 t/ha of fallen plant material. The mean harvest index (ratio of grain dry weight to total plant dry weight) of these cultivars was 0.24 excluding fallen material and 0.17 taking fallen material into account. Starch reserves were present in the stems during the vegetative phase, but disappeared during the reproductive phase. In 1974 the maximum leaf-area index on Vertisol was 3 and on Alfisol 12.7. The net assimilation rate tended to decline throughout the growth period, but in the medium cultivars increased at the end of the reproductive phase, probably because of photosynthesis in pods walls and stems.

In 1974 and 1975 the growth of roots and distribution of nodules in Vertisol was investigated by means of soil cores. Roots extended below 150 cm and root growth continued during the reproductive phase. Most nodules were found within the first 30 cm of soil, but some were found below 120 cm. In cv. T-21, grown in brick chambers 150 cm deep, at the time of harvest about three-quarters of the mass of the roots was found in the first 30 cm, and the shoot:root ratio was around 4:1.

In 1975 the mean uptake of nitrogen by the medium cultivars on Vertisol was 120 kg/ha, including 34 kg/ha in fallen material. In 1976 the uptake of nitrogen by cv. ICP-1 was 89 kg/ha on Vertisol and 79 kg/ha on Alfisol, including 32 and 23 kg/ha respectively in fallen material. Nitrogen uptake continued throughout the growing period. The percentage of nitrogen in stems and leaves declined as the plants developed and there was a net remobilization of nitrogen from these organs. The pattern of uptake and remobilization of phosphorus resembled that of nitrogen. In 1976 the total uptake of phosphorus by cv. ICP-1 on Vertisol was 5.8 kg/ha and on Alfisol 5.0 kg/ha.

The relatively low yields of pigeonpeas result from a restricted partitioning of dry matter into pods, which may be related to the plants' perennial nature.

In pigeonpeas (Cajanus cajan), most flowers are shed without setting pods. Pod-set is reduced by shading, defoliation and the presence of already developing pods, probably because of the reduced availability of assimilates or other nutrients. In pigeonpeas, unlike most leguminous crops, the average weight per pod of earlier and later formed pods is the same; this indicates that pod-filling is not limited by nutrient supply. Pod-set seems to be controlled in such a way that fewer pods develop than the plants are capable of filling. These processes can be represented by a simple working model, in which the assimilate supply corresponds to water in a reservoir, the axis of a branch or a raceme to a horizontal tube connected to the reservoir, and pods to containers of limited volume at a lower level; the connecting tubes between the axis and the 'pods' have an ascending limb, shorter than the descending limb to the pods, creating a siphon. 'Pods' can 'set' only when the level of water in the reservoir is higher than the threshold of the siphon; during the filling of earlier-set 'pods', the setting of other 'pods' is inhibited by the reduction of pressure within the axis. This model may provide a crude representation of mass flow within the phloem from sources to sinks; it also illustrates some of the hydrodynamical factors involved in competition among sinks.

Pigeonpeas (Cajanus cajan) are normally sown in June or July in India, at the beginning of the monsoon, but trials were carried out at Hyderabad by sowing in October or November as a winter crop. The duration of the crop, especially of the *medium* and *late* cultivars, was much reduced. In 1975*76, October-sown pigeonpeas gave yields comparable to those of the normal season but much lower yields were produced by planting in November 1975. *Medium* and *late* cultivars significantly outyielded early ones. Optimum plant populations for winter crops were 3*5 times higher than are normally used in the monsoon. Pigeonpeas at relatively high population densities could have considerable potential as a winter crop in peninsular India.

The number and percentage of nodes bearing two pods in 'double-podded' cultivars of chickpeas growth in northern India (at Hissar) and peninsular India (at Hyderabad) were compared. At Hissar 11% of the pod-bearing nodes were double-podded; at Hyderabad 28% were double-podded on early-sown and 49% on late-sown plants. In all cases the number of double-podded nodes per plant was similar, but different numbers of single- podded nodes per plant were formed, depending on the length of the growing season. At Hyderabad the percentage of double-podded nodes was not significantly affected by population-density nor by shading the plants throughout the reproductive phase. Partial defoliation of the plants reduced the percentage of double-podded nodes, as did the removal of all flowers from the plants for the first two to four weeks of the reproductive phase. The conversion of 'double-podded' plants to 'single-podded' plants by cutting off one of the flowers at every double-flowered node had no effect on yield at a location in the Himalayas where the double-podded character was poorly expressed, but at Hyderabad the yield of the 'single-podded' plants was significantly reduced compared with the 'double-podded' controls. The results indicate that the double-podded character can confer an advantage in yield of about 6 to 11% under conditions in which the character is well-expressed.

The symptoms of the pigeonpea wilt (causal fungus: Pusarium udum) generally appear during the reproductive phase, particularly while pod-filling is taking place (Mundkur, 1935).

In an off-season crop planted in December 1974 we observed that while there was a high incidence of wilt during the pod-filling phase of untreated plants, almost all the plants where pod development had been prevented by the removal of flowers remained healthy.

Conversely, we found that the incidence of the disease increased when the plants were defoliated during the reproductive phase. In an experiment carried out on medium- duration cultivars grown during the normal season (planted in June 1975) leaves were removed at the time flowering began, and subsequent defoliations were made as new leaves were produced. Different degrees of defoliation were employed: 33% (one leaf out of three removed), 50% (alternate leaves removed), 67% (two leaves out of three removed), 75% (three leaves out of four) and 100% (all leaves removed). We found that, in general, the incidence of the wilt increased with the severity of defoliation.

A second experiment was carried out on medium-duration plants (56 lines in the breeders' plots) which had been ratooned at the time of the harvest of the first flush of pods. These plants regenerated new branches and entered into a second reproductive phase, during which (on March 1 1976) one row of plants of each line was completely defoliated and another row was left as a control. Two months later the plants were scored for wilt. Of the controls, 16 out of 380 plants (4%) had wilted whereas 174 out of 360 defoliated plants (48%) had wilted.

Defoliation of plants in the ICRISAT patholigists' wilt-sick plot has also been found to lead to an increase in the incidence of the wilt disease.