Seeds of two chickpea cultivars; namely, Pirooz (desi-type) and Arman (Kabuli-type) were obtained from ICARDA. Seeds were surface sterilized in 1% sodium hypochlorite (NaOCl) for 3 min, subsequently washed with deionised water and air-dried before being used in experiments to avoid fungus attack. To study the effect of water stress on seed germination and seedling growth a wide range of water potential of PEG-8000 solutions: 0, −0.25, −0.5, −0.75, −1.0, −1.25, −1.5 MPa were prepared according to the procedure described by [13] . PEG- infused water agar Petri-dishes systems were used to impose precisely defined, constant, low water potential treatments. Water potential was lowered by the addition of various amounts of polyethylene glycol (PEG) (molecular weight 8000; Sigma, St Louis, MO) to the water-agar medium. Approximately 20 ml of PEG solution was poured on top of an equal volume of solidified water-agar in a Petri dish, and after 24 h, the solution on top of the dish was poured off and the dish used for experiments. During the 24 h, the PEG diffused into the agar medium, thus lowering its water potential. The 24 h period was long enough to reach an approximate equilibrium. The medium without PEG was poured in the Petri dishes for the control medium, which gave a water potential for the agar medium of approximately −0.25 MPa [13] [14] . The experiment was conducted at different temperatures: 10˚C, 15˚C, 20˚C, 25˚C, 30˚C, and 35˚C in darkness.

The treatments were placed in a factorial arrangement in a completely randomized design with four replications of 25 seeds each. The number of germinated seeds was counted every 2 days. A seed was considered germinated when the emerging radicle elongated to 2 mm. Four characteristic of seed germination were determined: final germination percentage, mean time to germination (MTG), coefficient of variation of germination time (C.V%) and germination rate index (GRI).

MTG was estimated according to the following formula:

where:

ni = is the number of germinated seeds at day (i),

di = is the incubation period in days

N = is the total number of germinated seeds in the treatment.

GRI was estimated according to the following formula:

where:

N = is the total number of germinated seeds in the treatment,

P = is the number of incubated seeds at the beginning of the experiment,

ni = is the number of germinated seeds at day (i),

di = is the incubation time in days.

All seeds from the previous germination tests which did not germinate after 40 days at different PEG 8000 solutions, were removed and placed in new sterilized Petri dishes with autoclaved filter paper moistened with deionized water, and incubated under the same conditions for additional 40 days to study the recovery of germination. Recovery, for each specific osmotic environment, was calculated as the percentage of ingeminated seeds after PEG treatment to that germinated in the subsequent germination test in deionized water. Germination and recovery percentages were arcsine-transformed before statistical analysis to ensure homogeneity of variance. Data were analyzed using SPSS for windows, version 15. Analysis of variance (ANOVA) was performed to detect differences in all parameters studied followed by least significant difference test (LSD) to estimate least significant range between means.

The main effects of cultivar, temperature and water potential and their interactions were significant (p < 0.05) for all investigated characters. Germination was positively affected by increasing temperature up to 25˚C and thereafter there were a considerable decrease at 30˚C and 35˚C (Table 1, Figure 1). Supporting evidence was

^*The same lowercase letter (s) for temperature in the row and the same uppercase letter (s) for water potential in the column are not significantly different at 5% probability.

reported by many researchers [7] [8] [15] [16] . For both cultivars, the final germination percentages and rate of germination were highest at 25˚C and then started to decrease as temperature increased regardless of water potential (Table 1, Figure 1). However, both cultivars showed the lowest germination percentage and rate of germination at 10˚C under all water potentials, and at this temperature no cultivar was able to germinate at water potential less than −0.75 MPa (Table 1, Figure 1). Similarly, as water potential decreased, the final germination and germination rate index (GRI) were significantly decreased in both cultivars at all temperatures (Table 1, Figure 1). These results may be attributed to the fact that large seeded species such as chickpea need more thermal requirements to start germination. The sensitivity of seed germination and seedling growth to low water potential at high temperatures (30˚C and 35˚C) may be due to the physiological adjustment of seeds to conditions near thermal and/or water potential threshold as reported by many researchers [17] [18] . According to [19] , temperature requirement for seed species in arid and semi-arid regions to achieve 60% - 100% germination ranges from 15˚C - 35˚C, with temperature between 20˚C - 25˚C being suitable for most species. Moreover, the result showed that the effect of temperature on germination was much greater than water stress. Some researcher such as [20] indicated that since temperature had a strong influence on germination, seeding time should be selected to match the expected temperature required for successful germination at any particular location. They added that early high germination across a large range of temperature might be beneficial for rapid establishment of the crop in semi-arid warm regions where soil moisture in the upper soil surface is available for only short period. The effect of low water potential was pronounced on seed germination of chickpea at different temperatures.

For the two cultivars studied there were a threshold temperature and water potential under which germination does not occur. Similar results were reported by other researcher [7] [8] [14] - [16] [21] . Seed germination of chickpea under suboptimal conditions, simulated by different temperature or water potential in this study, is strongly influenced by genotype (cultivar). The cultivar Arman tended to germinate faster than cultivar Pirooz regardless of temperature and water potential.

The genotypic differences in response to temperature and water potential for final germination percentage and GRI were highly significant. In this regard, the cultivar Arman exhibited higher final germination percentage and rate of germination compared to cultivar Pirooz under all treatments (Table 1, Figure 1).

In contrast to germination percentage and germination rate index, the cultivar Arman had lower MTG compared with the other cultivar Pirooz regardless of temperatures and water potentials (Table 2). Decreasing water potential resulted in substantial delay in MTG for both cultivars at all temperatures studied (Table 2). On the other

^*The same lowercase letter (s) for temperature in the row and the same uppercase letter (s) for water potential in the column are not significantly different at 5% probability.

hand, increasing temperature up to 25˚C had a positive effect on MTG (both cultivars took less time to 50% germination). However, high temperatures (30˚C and 35˚C) resulted in remarkable delay of germination, in both cultivars, regardless of water potential (Table 2). The present study revealed that seed germination of chickpea under suboptimal conditions, simulated by different temperature or water potential in this study, is strongly influenced by genotype (cultivar). The cultivar Arman tended to germinate faster than cultivar Pirooz regardless of temperature and water potential. Moreover, the cultivar Arman had considerably higher recovery percentage from water stress compared with Pirooz. Thus, it obvious that cultivar Arman is more tolerant to water stress than the cultivar Pirooz. The cultivar Arman may have better seedling establishment and increased survival ability than the cultivar Pirooz under water stress conditions. This might be attributed to the fact that the roots of the cultivar Arman could elongate rapidly even under unfavorable soil moisture, thus ensuring the continuation of water supply to the plant.

For both cultivars, transfer of non-germinated seeds from PEG solution to the deionized water resulted in substantial increase in recovery percentage at all temperatures except at 10˚C, at which no germination occurred (Figure 2). The results revealed that the cultivar Arman scored the highest recovery percentage from water stress at all studied temperature compared with the other cultivar Pirooz. Surprisingly, the results also indicated that temperature had a remarkable effect on recovery from water stress (Figure 2). In this respect, increasing temperature significantly increased the recovery percentage up to 25˚C, with substantial decrease in recovery percentage at high (30˚C and 35˚C) temperatures (Figure 2). According to these findings, both chickpea cultivars have tolerance ability to water stress as their seeds resumed germination and recover from water stress. The temperature seems to have major effect on recovery from stress as there is an increase in the recovery with increasing temperature regardless of water potential. The study carried out by [14] [22] [23] indicates the significance of thermo-period in determining the recovery of germination responses of halophyte seeds.