Water scarcity is a production problem faced by most peanut growers worldwide. Peanut (Arachis hypogaea L.) originates from South America (Hammons, 1982) and is adapted to warm temperatures and long growing seasons. Therefore, peanut is grown almost exclusively in regions prone to extended, or at least partial, periods of drought during the growing season. Irrigation has proven to be a boon to producers, increasing yields by as much as 19% (Lamb et al., 1997), but is often either severely limited or non-existent for many growers. Therefore, research in peanut production needs to be conducted concerning: 1) the sustainable use of limited irrigation; and 2) the development of genotypes that are best adapted to these conditions, particularly those peanut genotypes that are highly water use efficient. However, increased water use efficiency must accompany the maintenance of high yield to have any utility in agricultural systems. The intrinsic physiology of peanut matches this criteria; peanut has the potential to have very high photosynthetic capacity accompanied by low stomatal conductance levels, translating into high water use efficiency without sacrificing carbon assimilation and possibly yield (Wright et al., 1993).

Physiological water use efficiency (WUE) is defined as the ratio of photosynthesis to transpiration and is often a limitation to peanut productivity under drought (Nageswara Rao and Wright, 1994). However, determining the variability in WUE among peanut genotypes or pinpointing the production conditions that significantly affect WUE can be extremely problematic due to the complexity and tedious nature of direct measures of WUE (Wright et al., 1993). Traditional procedures of measuring WUE require weighing lysimeters and accurate measurements of water applied throughout the growing season (Wright et al., 1988; Hatfield et al., 1989). In the last two decades, extensive research has identified an accurate surrogate for WUE in peanut through the measurement of natural δ¹³C composition in plant tissue. The isotopic discrimination of ¹³C that occurs during the photosynthetic pathway in C₃ plants (Farquhar et al., 1982; Farquhar and Richards, 1984) is well correlated with WUE in peanut and provides a long-term measurement of WUE across the season (Wright et al., 1993; Nageswara Rao and Wright, 1994). Water-use efficiency is positively correlated with δ¹³C composition (i.e. the higher the δ¹³C composition, the higher the WUE).

While extensive information documenting varietal differences in δ¹³C content (and thus WUE) have been conducted in Australia, almost nothing is known about the variability in isotopic composition among genotypes utilized in U.S. peanut producing regions. Surveys of the genetic variation in δ¹³C are needed for U.S. peanut genotypes, and the role that environment and production practices, particularly irrigation, play in changing the expression of these genetic differences needs to be quantified. This type of information is invaluable for developing more water use efficient varieties adapted to the U.S. production environment.

Additional information is needed about the interaction of δ¹³C with other leaf phenotypic characteristics, particularly δ¹⁵N. Theoretical analyses and field data suggest that WUE may interact with nitrogen fixation and other nitrogen nutrition effects (Schulze et al., 1991; Guehl et al., 1998). The overriding pattern appears to be a positive correlation between δ¹³C and δ¹⁵N, such that N₂ fixation (low δ¹⁵N values) is associated with reduced water-use efficiency (more negative δ¹³C) (Schulze et al., 1991; Handley et al., 1994; Knight et al., 1993). Therefore, in studies examining δ¹³C in a leguminous crop, it is important to likewise examine the relationship between δ¹³C and δ¹⁵N and the implications for nitrogen fixation. Lastly, the relationship of δ¹³C with the leaf characteristics SPAD chlorophyll content and specific leaf area (SLA) would be important to evaluate because the characteristics are being used in other breeding programs and could potentially be screening tools for WUE (Nageswara Rao and Wright, 1994; Nageswara Rao et al., 1995). None of this information is available for U.S. peanut cultivars or developing breeding lines.

This experiment was conducted in order to document genetic variation and the effect of irrigation environment on δ¹³C, δ¹⁵N, and other leaf characteristics in U.S. grown peanut genotypes and to quantify the relationship between δ¹³C, SLA, SPAD, and δ¹⁵N. The specific experimental objectives were: 1) to determine if genetic variability existed in δ¹³C, and δ¹⁵N SLA, SPAD among three commonly grown U.S. peanut genotypes; 2) to determine if differing irrigation levels affected the pattern of variability; and 3) to quantify the relationship between δ¹³C and several leaf phenotypic characteristics.

Total water received within each water treatment ranged from 711 and 627 mm ha⁻¹ in 2001 and 2002, respectively, for the 100% irrigation treatment and 528 and 43 mm ha⁻¹ in 2001 and 2002, respectively, for the non-irrigated treatment (Table 1). Water treatment affected peanut yield, with no significant differences between 2001 and 2002. Across years, there were significant differences in yield for peanut genotype (df  =  2, F Ratio  =  5.9, P value  =  0.0048), with Georgia Green having significantly higher yields than the other two genotypes. In addition, irrigation treatment significantly affected yield across genotypes (df  =  3, F Ratio  =  18.4, P value  =  0.0001) such that the NI treatment yields were significantly lower than any of the irrigated treatments (33%, 66%, or 100%; Table 2). In both years the yield in the 100% treatment was numerically lower than the 66% treatment, indicating possible over-watering in the former treatment.

Significant differences in δ¹³C, δ¹⁵N, SPAD chlorophyll content, SLA, %C, %N, and C to N ratio were found between years, and differences existed in all traits except δ¹⁵N among genotypes (Table 3). Irrigation effects were less apparent with differences existing only for δ¹⁵N, SPAD chlorophyll, and SLA. Year by genotype interactions were significant for both δ¹³C and δ¹⁵N, while year by irrigation interactions were significant for δ¹⁵N, SPAD, specific leaf area, %N, and C to N ratio. There was no significant genotype by irrigation effect. The three-way interaction among year, genotype, and irrigation was significant only for %C (Table 3).

Mean δ¹³C was significantly lower (more negative) in 2002 than in 2001; while mean δ¹⁵N, specific leaf area, and %N were greater in 2002 (Table 4). Multiple comparisons tests revealed that the significant differences among genotypes for all traits except δ¹⁵N were due to the cultivar Georgia Green having significantly lower δ¹³C and SPAD chlorophyll content than the other two genotypes and lower %N than the cultivar C99R; while SLA and %C for Georgia Green were significantly greater than for the cultivar AT201. The significant effects of irrigation treatment revealed higher values of δ¹⁵N and SLA in the NI treatment as compared to the other three irrigation levels, and higher SPAD chlorophyll content in the NI treatment in comparison to the 100% irrigation level. For mean SLA values, the 66% irrigated treatment had the lowest SLA in comparison to the other three treatments (Table 4).

The relationship between δ¹³C and peanut yield (across irrigation treatments) is shown in Figure 1 for 2001 and 2002. In 2001, the correlation between δ¹³C and peanut yield was only significant for the cultivar C99R, which showed a negative correlation with yield as δ¹³C increased (became less negative). In 2002, yield data on C99R was unavailable; however, the correlation between δ¹³C and yield was significant for the cultivar AT201, but in this case, was a positive relationship such that δ¹³C increased (became less negative) with increasing yield. There were limited significant correlations with yield for other phenotypic traits (Table 5). For the cultivar AT201, SPAD chlorophyll content in 2001 and 2002 was significantly negatively correlated. None of the other cultivars showed such a relationship. Both C99R and Georgia Green had significant negative correlations of yield with SLA in 2001 and 2002 (with the exception of no measured yield for C99R in 2002). C99R further showed a negative relationship with yield for δ¹⁵N; while Georgia Green had a negative and positive relationship with yield for %N and C/N, respectively.

The phenotypic correlations between δ¹³C and the other measured traits were equally sparse (Table 5). In 2001, significant positive relationships between δ¹³C and SLA for C99R and SPAD chlorophyll for Georgia Green were noted. For Georgia Green in this same year, %C was negatively associated with δ¹³C. In 2002, SPAD was negatively correlated with δ¹³C for AT201, while Georgia Green showed a positive correlation between δ¹³C and δ¹⁵N. There were no significant correlations between δ¹³C and the other phenotypic traits for the cultivar C99R in this same year.

The drought susceptibility index (S) showed interesting differences among cultivars for the various phenotypic traits in both 2001 and 2002 (Figure 2). Overall and with few exceptions, the cultivar AT201 exhibited an advantage (lower values) in the drought susceptibility index in comparison to the other genotypes for δ¹³C, δ¹⁵N, SLA, and SPAD for both years. On the other hand, the cultivar Georgia Green (with the exception of its S value for δ¹⁵N in 2001) exhibited a disadvantage in drought tolerance with its consistently higher S values for the four traits measured. The cultivar C99R primarily exhibited intermediate S values between the cultivars AT201 and Georgia Green for the four traits.

Significant variation in isotopic composition and related leaf phenotypic characteristics was found for three Arachis genotypes that are currently utilized in the southeastern U.S. production regions. This finding supports the possibility of selecting for U.S. genotypes with increased δ¹³C (and thus WUE) in programs aimed at increasing the efficiency of peanut production under water scarce environments. These results also concur with previous studies that found genotypic variation in δ¹³C for peanut in greenhouse (Hubick et al., 1986) and field conditions (Nageswara Rao et al., 1993). In peanut, variation in δ¹³C appears to be due to variability among genotypes in their photosynthetic capacity and not stomatal factors (Nageswara Rao et al., 1995; Nageswara Rao et al., 2001). The δ¹³C content of leaf tissue can be controlled either through an intrinsic photosynthetic capacity (capacitance types) or through stomatal diffusive properties (conductance types) (Udayakumar et al., 1998). In conductance type plants, high δ¹³C concentrations, and thus high WUE, is at the expense of dry matter production and yield. In contrast, peanut is a capacitance type plant that has the ability to concentrate δ¹³C in its tissues without stomatal closure, thereby making it possible to select for high δ¹³C and high WUE while maintaining high growth rates and yield (Udayakumar et al., 1998). Mean differences among genotypes in this study revealed that the cv. Georgia Green had significantly lower δ¹³C, and thus WUE, in comparison to the other two genotypes.

While the current results agree with findings of variability in δ¹³C among peanut genotypes, they counter many studies examining irrigation effects on δ¹³C in other crop species. What was surprising in this study was the lack of effect of irrigation on the expression of δ¹³C, which is contrary to other studies in crops such as wheat, cotton, (Yakir et al., 1990; Saranga et al., 1998; Monneveux et al., 2005; Merah et al., 1999) and even loblolly pine (Choi et al., 2005). In these studies, δ¹³C was increased under non-irrigated or drought stressed conditions, indicating that WUE increases under water scarcity. In the current study, the only significant environmental effects on δ¹³C were reflected through the significant effect of year, indicating the climatic variation between 2001 and 2002 had an effect on δ¹³C. Further, the existence of a significant G X E interaction between genotype and year for δ¹³C (Table 3) indicated that the genetic effects on δ¹³C were strong, but environment played a role as well. Evidence for G X E interactions for WUE in peanut is scarce (Hubick, 1990; Nageswara Rao and Wright, 1994). A few previous studies in peanut documented that variation in δ¹³C was influenced by location and genotype (Nageswara Rao and Wright, 1994; Brown and Byrd, 1996); while Wright et al. (1988) found strong genetic control over δ¹³C with little effect of the environment in four peanut genotypes when grown either in open or closed canopies.

There were significant genetic and environmental effects on the other phenotypic traits measured. Genetic variation was strong for all the leaf characteristics except δ¹⁵N; in addition, significant variation among years was present indicating an environmental component to control of the expression of these traits. But in this instance, irrigation environment also had a significant role in determining the variability in δ¹⁵N, SPAD, and SLA. Irrigation, in general, tended to decrease δ¹⁵N, SPAD, and SLA. For δ¹⁵N, these results are in line with the study of Handley et al. (1999) that found the natural abundance of δ¹⁵N in leaf tissue decreased with increasing soil moisture and therefore, was reflective of water availability at a given site. Our results for SLA, however, are contrary to most studies. Decreased SLA under limited irrigation is a common finding in many drought studies (Marcelis et al., 1998), and may be a mechanism of increasing water use efficiency through the concentration of chlorophyll and proteins in thick leaves leading to greater photosynthetic capacity (Liu and Stutzel, 2004). The SPAD chlorophyll content results in this study do show that the highest chlorophyll concentrations are in leaves in the NI treatment, but are concentrated in thinner leaves in comparison to the irrigated treatments. Given the low SLA in the irrigated treatments and the lack of irrigation effect on δ¹³C, severe drought stress was likely not present in either 2001 or 2002.

The relationship between δ¹³C and yield was complicated for the genotypes tested and appeared to vary among the cultivars. The cultivar Georgia Green showed no significant correlation with δ¹³C and yield in either 2001 or 2002. A reduced correlation between δ¹³C and yield in wheat has been linked to an absence of variability in δ¹³C. This can be explained by increased photosynthetic capacity being offset by increases in stomatal aperture, leading to lower variation in δ¹³C and reduced correlation between δ¹³C and yield (Monneveux et al., 2005). This may indicate that Georgia Green tends to be a conductance physiological type instead of a capacitance type. The high relative drought susceptibility index values (S) for Georgia Green for the trait δ¹³C also indicate that this genotype is more drought susceptible, and indeed its mean δ¹³C values signify lower WUE than the other genotypes. The cultivar C99R showed a negative correlation with δ¹³C and yield, indicating that this genotype may also be acting as a conductance type, where δ¹³C (and thus WUE) is controlled predominantly through stomatal aperture. In the case of C99R, WUE increases only through stomatal closure and so an increase in WUE brings a concomitant decrease in yield. On the other hand, the genotype AT201 appears to be acting as a capacitance type where increased δ¹³C (and thus WUE) is correlated positively with yield in 2004, likely due to an increase in photosynthetic capacity. The low drought tolerance index values for δ¹³C also indicate that AT201 may be more water use efficient than the other genotypes, possibly through increased photosynthetic capacity.

The relationships between δ¹³C and yield with the other leaf phenotypic characters were extremely limited. Very few traits correlated directly with yield, with the exception of nitrogen related traits (δ¹⁵N, and %N) for Georgia Green. Specific leaf area was the most consistent correlation with yield for genotypes and years, indicating that thicker leaves were correlated with higher yields. The relationships with δ¹³C and leaf phenotypic characters were nearly nonexistent. In 2002, the relationship between δ¹⁵N with δ¹³C was significantly positive for Georgia Green. Very few studies have related δ¹⁵N with δ¹³C, and those that have, show the relationship to be affected by environment or nonexistent (Guehl et al., 1998). The paucity of significant correlations between δ¹³C and SPAD or SLA is unfortunate. In Australia and India, research has confirmed the relationship between SLA and SPAD, two very easily measured characters, with carbon isotope discrimination (Wright et al., 1993; Nageswara Rao and Wright, 1994; Nageswara Rao et al., 1995). This relationship has been purported as a useful, inexpensive, and easily utilized tool to screen large numbers of breeding lines for water use efficiency. However, this study has failed to show a consistent relationship between the two traits in a field setting for these three peanut cultivars. Nageswara Rao et al. (2001) also showed a link between SLA and SPAD chlorophyll content, and it was assumed that because δ¹³C and SLA are related, δ¹³C and SPAD would be related as well. However, in this study, the correlation between δ¹³C and SPAD was significant only for Georgia Green in 2001 and AT201 in 2002, with the direction of the correlation being inconsistent between cultivars. Overall, it appears that the utility of SLA and SPAD chlorophyll for WUE screening tools in U.S. peanut genotypes in a field environment is limited because of the weak correlations between δ¹³C and SLA or SPAD.

In conclusion, this study has documented variation in δ¹³C, δ¹⁵N, and other leaf phenotypic characteristics among currently grown U.S. peanut genotypes. Based on the copious number of studies documenting the plasticity of δ¹³C in other crop species, it is surprising that there was no significant effect of irrigation on the trait in this study. It might be argued that the crop received adequate water through precipitation regardless of supplemental irrigation level in both years, thus negating the effect of irrigation on δ¹³C. However, the differences in yields among the irrigation levels in both 2003 and 2004 somewhat refute this explanation. Perhaps the explanation lies in a relatively strong genetic control for δ¹³C in peanut as compared to other crops, making the success of breeding for increased WUE in peanut quite achievable.