GRAIN LEGUME SPECIES IN LOW RAINFALL MEDITERRANEAN-TYPE ENVIRONMENTS .2. CANOPY DEVELOPMENT, RADIATION INTERCEPTION, AND DRY-MATTER PRODUCTION

The adaptation of a wide range of grain legume species (Lupinus albus L. cv. Kiev mutant, L. angustifolius L. cv. Yorrel, L. atlanticus L. a ccs. P22924 and P22927, L. pilosus Murr. acc. P23030, Cicer arietinum L. acc. T1587, Lens culinaris Med. acc. ILL6002 and cv. Digger, Vicia faba L. cv. Fiord, V. narbonensis L. acc. ACT60104, Lathyrus cicera L. acc. 495, L. ochrus (L.) DC acc. 537, L. sativus L. acc. 453, P. sati vum L. cv. Dundale, V. benghalensis L. cv. Early purple, and V. sativa L. cv. Languedoc) to low-rainfall Mediterranean-type environments of southwestern Australia was examined in relation to canopy development, radiation interception, and dry-matter production. Species were grown at one location, and in two consecutive growing seasons (1993 and 199 4). Dry-matter production was large for most species, particularly in 1993 (maximum biomass > 580 g/m(2) for all species except L. pilosus) which had above-average rainfall. V. faba and P. sativum developed gre ater maximum biomass than other species, and this was associated with high crop growth rates (CGR) during the early part of the growing seas on and large maximum CGR. Other species which produced large maximum b iomass (V. narbonensis and L. ochrus in 1993) developed dry matter mor e slowly during the early stages of plant growth, but achieved maximum CGR similar to those obtained for V. faba and P. sativum. CGR of thes e species increased rapidly close to flowering as air temperatures beg an to rise. Species which produced large maximum biomass intercepted a greater amount of photosynthetically active radiation (PAR) (r=0.73 and 0.70* in 1993 and 1994, respectively). There were differences amo ng species in the efficiency with which intercepted PAR was converted into biomass (radiation-use efficiency), but these differences did not closely reflect differences in dry-matter production. Cumulative inte rcepted PAR was positively correlated with the fraction of incident PA R intercepted by species (r=0.87* and 0.96** in 1993 and 1994, respec tively), but was poorly correlated with cumulative incident PAR. Diffe rences among species in intercepted PAR were therefore not related to differences in the length of time species were intercepting incident P AR (a function of crop phenology). Greater fraction of incident PAR in tercepted by species was closely related to green (photosynthetic) are a duration (GAD)(r=0.60 and 0.78* in 1993 and 1994, respectively), bu t was poorly correlated with the efficiency with which green area inte rcepted PAR (described by the extinction coefficient). Maximum biomass of species was also closely related to GAD (r=0.78* and 0.84** in 19 93 and 1994, respectively). These results indicate that in low-rainfal l Mediterranean-type environments, grain legume species should be sele cted for the development of large green area index (GAI), which should maximise the interception of PAR, production of dry matter, and conse quently seed yield of the crop. (C) 1997 Elsevier Science B.V.