The fruit plus accessory parts, and thus not only the fruit or seed, is the diaspore in some species of flowering plants. Atriplex centralasiatica, a summer annual salt-secreting cold-desert halophyte of central Asia that produces flat and humped diaspores, is such a species. The dispersal/germination units of this diaspore-heteromorphic species are fruits enclosed in persistent ‘bracteoles’. Germination of black and brown fruits (i.e. ‘bracteoles’ removed) of this species has been studied in some detail previously, but little attention has been given to the germination biology of the intact diaspores. The aim of this study was to compare the germination biology of the intact diaspores of A. centralasiatica, and their fates in the field during 12 months on the soil surface and buried at 5-cm depth. Fresh flat diaspores germinated to 42.7–51.3% in light (12-h photoperiod) and to 16.0–59.3% in constant dark, over a temperature range of 15–20/30°C, while no freshly matured humped diaspores did so under any temperature/light regime. Neither gibberellin (GA3) nor potassium nitrate (KNO3) had an effect on germination of either diaspore. Both diaspores readily imbibed water, and removal of ‘bracteoles’ released most of the dormancy. Thus, the ‘bracteoles’ are primarily responsible for diaspore dormancy in A. centralasiatica. Humped diaspores persisted for a longer period in the soil than flat ones. The germination requirements of intact natural dispersal/germination units of A. centralasiatica differ from those previously reported for fruits of this species.

An in vivo digestibility trial was conducted by feeding sheep the leaves, fruits and twigs of Atriplex barclayana in a proportion roughly equivalent to that eaten by sheep grazing freely in Atriplex plantations. Four treatments were imposed on each of four sheep in a 4 × 4 Latin-square experimental design: Atriplex offered alone or with 100, 200 or 300 g/day tapioca meal.

The mean apparent digestibility of the Atriplex dry matter (DM) and organic matter (OM) consumed were 0·59 and 0·56, respectively. Addition of tapioca to the Atriplex in the diet did not improve these digestibility coefficients. The low OM content of 760 g/kg together with its digestibility of 0·56 resulted in the Atriplex having a low metabolizable energy concentration of 6·28 MJ per kg DM. The in vitro apparent DM digestibility of Atriplex was approximately 0·09 higher than the in vivo apparent digestibility.

The mean nitrogen concentration of the Atriplex DM was 16·6 g/kg, and its apparent digestibility 0·73, which was not improved by the addition of tapioca to the diet. Nitrogen retention of the sheep eating only Atriplex was proportionately 0·17 of the nitrogen intake. The addition of 300 g tapioca improved nitrogen retention to 0·27 but was not significantly different from the other treatments.

Water intake and urine excreted were as high as 14 and 12 I/day respectively, for an Atriplex DM intake of about 1300 g/day. During the experiment the sheep only maintained live weight, despite daily intakes of up to 1200 g Atriplex DM and up to 300 g tapioca.