Because of limited seed and inconsistent germination of most wild Arachis accessions, they were vegetatively propagated from May to September 2019 to obtain plants for the experiments. Mother plants were maintained in pots (33-cm-top diam., 20-cm-bottom diam., 15-cm tall) filled with the soilless mix BM7-35 (Berger). Greenhouse temperatures were set at 22 to 32 C. A small amount of Rhizobium inoculant (Guard-N; Verdesian Life Sciences) was applied by dipping the seed into the inoculant powder immediately before planting or by placing ca. 3 mm³ of the inoculant into the planting hole.

Cuttings were taken from branches with 3 to 4 nodes, and leaves, pegs, and flowers were removed from the lowermost 1 to 2 nodes. The bottom nodes were dipped in rooting powder (Garden Safe TakeRoot; Spectrum Brands Holdings) before placing in 9.5-diam. pots filled with a soilless propagation mix (Sunshine Redi-Earth Plug and Seedling; Sungro). Rhizobium inoculant was applied to the planting hole as previously described before cuttings were inserted into the potting mix. Potted cuttings were placed in an 0.9-m-wide x 3.3-m-long x 0.9-m-high enclosed humidity chamber constructed from 2.54-diam. polyvinyl chloride piping and 6-mil Solar-Ice polyethylene film (now Luminance, RPC BPI Agriculture). The humidity chamber was partially covered with 50% Aluminet (Ecologic Technologies, Inc.) to further reduce heat from infrared radiation. Overhead misting lines (Orbit Irrigation Products) provided high humidity by misting approximately 5 sec every 5 min. After four weeks, cuttings were monitored weekly for root development by gently pulling on the stem. Rooted cuttings were removed from the chamber and placed in shade for 2-3 weeks to acclimate to lower humidity. Cuttings were then transplanted into 15-cm-diam. pots filled with BM7-35 soilless mix and moved into full light on greenhouse benches. A 46-cm-long x 6.35-diam. acrylic rod (McMaster-Carr) was inserted into each pot, and top-heavy and trailing plants were tied to the rod with paper wire twist ties. In order to prolong the life of biennial and annual species (Kvien and Ozias-Akins, 1991), the vegetatively propagated plants were monitored weekly to remove any pegs that had developed. Plants were also pruned as needed to keep branches from extending ca. 30 cm beyond the top of the acrylic rod.

For the resistant and susceptible A. hypogaea reference genotypes, three seeds were planted with Rhizobium inoculant in 15-cm-diam pots filled with the BM7-35 potting mix. Pots were planted eight weeks before inoculation and thinned to one plant after emergence. All wild and cultivated plants were fertilized with 15 mL of NPK 14-14-14 slow-release product (Osmocote Smart Release Flower and Vegetable; ScottsMiracle-Gro). In addition, micronutrients were applied by adding Fertileader Vital (300 mL/379 L water; Timac Agro USA) to the irrigation water until a minimum of 2 weeks before inoculation.

Athelia rolfsii mycelial plugs were prepared and plant inoculations were conducted as described in a previous study (Bennett, 2020) with the exception that plants were inoculated on the main stem, usually at the 2^nd node, approximately 30-60 mm from the soil line. Because A. microsperma PI 674407 had a particularly bushy growth habit, the impeding branches were tied upwards with string to facilitate inoculation and data collection. Plants were placed inside the humidity chamber, where high humidity was maintained for part of the day with an ultrasonic transducer fogging system on each end of the chamber. The fogging system, hooked up to a water hose, was constructed using a 39 cm x 30 cm x 17 cm project box, a water tank float to maintain proper water level, a 24 V computer fan, four 24 V ultrasonic transducers, and a 24 V power supply (Figure 1). Each ultrasonic transducer was able to diffuse between 200-400 mL of water per hour. The fogging systems operated 9 hr per day, with one running between 0600 and 1500 hr, and the other between 1100 and 2000 hr. Temperature and relative humidity were monitored each hr with three HOBO U23 data loggers (Onset Computer Corp.) starting 25 Feb. 2020. Main stem lesion length was measured 4, 6, 10, and 12 d after inoculation using a digital caliper (Mitutoyo America). Because mycelium frequently obscured lesions, it is often faster to measure mycelium than to measure lesions. Thus, mycelium measurements were taken to see if they could be substituted for lesion measurements. If the plant died before the end of the experiment, measurements for remaining days were recorded as missing data. The experiment (trial) was conducted 11 times between 9 Feb. 2020 to 29 Apr. 2020.

Plant pots were arranged inside the chamber in a randomized complete block design with two replications. Data were analyzed using SAS Version 9.4 (SAS Institute). Differences among entries in lesion length and mycelial growth were determined using repeated measures ANOVA in PROC MIXED with TOEP covariance structure. Trial and block(trial) were used as random variables in the model. Differences among and within entries at 4, 6, 10, and 12 d after inoculation were analyzed using the SLICE option. Correlation analysis between lesion and mycelium lengths were conducted using PROC CORR. Area under the disease progress curves (AUDPCs; Shaner and Finney, 1977) for lesion length were estimated and analysed using PROC GLIMMIX. AUDPC means were compared using a split-plot design with trial as the whole plot and entry as the subplot, and the SLICE option was used to examine differences among trials and entries. All pairwise comparisons were adjusted for Type I error with the ADJUST = TUKEY option at α = 0.05.

In the repeated measures analyses of lesion length, the interaction between entry and time was significant, indicating that differences among entries depended on the day of observation (F = 1.44; df = 51, 827; P = 0.03). Lesion lengths among entries differed on d 6, 10, and 12 (Table 2) but not on d 4 (F = 0.63; df = 22, 693.1; P = 0.91). On d 12, the longest lesions (> 50 mm) were found on the two A. batizocoi accessions, and A. kuhlmannii (Table 2). The shortest lesions (≤ 23 mm) were found on A. cardenasii, A diogoi, and A. microsperma PI 674407 and PI 666096. When sliced by entry, lesions lengths were significantly different over time for all entries except A. correntina, and the two A. microsperma PIs (Table 2).

The interaction between entry and time (F = 1.41; df = 51, 834; P < 0.03) was also significant for mycelial growth. Significant differences among entries were found for all days except d 4 (F = 0.53; df = 17, 541.3; P = 0.94). Mycelial growth occasionally decreased on some plants between d 10 and 12, but entries with the most mycelium (> 52 mm) for both days were A. kempff-mercadoi, A. hoehnei, A cruziana, A. monticola, and A. batizocoi PI 468327 (Table 2). These entries sustained significantly more mycelial growth than both A. microsperma entries by d 12. The correlation between mycelial growth and lesion length over time was considerable (r = 0.74; P < 0.01), but not as strong as in a previous study consisting only of A. hypogaea entries (r = 0.92; P < 0.01; Bennett, 2020). When the relationship was examined by day, the best correlation was found on d 10 (r = 0.76; P < 0.01), followed by d 12 (r = 0.73; P < 0.01), d 6 (r = 0.52; P < 0.01), and d 4 (r = 0.36; P < 0.01). Most entries had strong correlations (r > 0.70) between mycelial growth and lesion length except A. monticola (r = 0.40; P < 0.01), A. cruziana (r = 0.61; P < 0.01), A. diogoi (r = 0.62; P < 0.01), and A. kempff-mercadoi (r = 0.69; P < 0.01). Interestingly, A. monticola supported substantial mycelial growth but had comparatively smaller lesions on d 10 and 12. Thus, it appears that mycelial growth may not be a suitable indicator of resistance for comparisons among Arachis species.

Area under the disease progress curves were also examined. Missing data for the 11 entries with fewer plants resulted in non-convergence of the model, so data from all entries in trials 1 to 5 were used in one analysis, and data from the seven entries used in all trials were analyzed separately (Table 3). For both analyses, the effects of trial (trials 1-5, P = 0.21; trials 1-11, P = 0.41) and trial x entry (P > 0.24) were not significant. However, the effect of entry was significant (P < 0.01). In the first five trials, the highest AUDPC was in A cruziana, followed by the two A. batizocoi accessions; the two A. microsperma accessions had the lowest AUDPCs. When the seven entries were compared over all trials, A. kuhlmannii had the highest mean AUDPC, which was significantly greater than those of A. cardenasii, A. diogoi, and both A. microsperma PIs. To note, A. simpsonii was among the more resistant entries in trials 1-5 but appeared considerably more susceptible in the analyses of all trials. Similarly, the AUDPC of A. kuhlmannii increased by 62 mm when the later trials were added. These changes in susceptibility may be due to the higher temperatures observed in the later trials, and A. simpsonii susceptibility to At. rolfsii may be particularly temperature sensitive. Pande et al. (1994) also observed that plant mortality to At. rolfsii increased with higher temperatures in some genotypes.

Compared to a previous study (Bennett, 2020), the greenhouse humidity chamber used here resulted in less severe disease, especially among the susceptible control genotypes CC038 and CC041. The difference in results is likely due to the constant optimum temperature maintained by growth chambers in that study. However, CC038 and CC041 were significantly more susceptible to A. rolfsii than the resistant genotypes Georgia-03L and CC650 (Table 2), consistent with recent laboratory and field studies (Bennett, 2020; Bennett and Chamberlin, 2020). The susceptible cultivar NC-V11 was numerically intermediate in lesion length and mycelial growth to the susceptible and resistant genotypes and differed statistically only from CC650 in lesion length on d 12. Both the greenhouse and growth chamber assays were better at identifying highly susceptible and some highly resistant genotypes than discriminating among intermediate entries (Bennett, 2020). Despite this limitation, laboratory assays are less likely to be influenced by canopy microclimates than field evaluations because high levels of humidity are easier to maintain. Arachis spp. vary considerably in canopy architecture, e.g. ranging from the relatively short and ramose A. microsperma to the little-branched, 1-m-tall mainstem (and up to 4-m-long lateral branches) of A. batizocoi (Krapovickas and Gregory, 2007).

Comparisons between the reference A. hypogaea genotypes and wild species were not made due to differences in plant age at inoculation. Several studies have observed that susceptibility to At. rolfsii decreases with plant age (Pande et al., 1994; Pratt and Rowe, 2002; Bekriwala et al., 2016), yet the younger A. hypogaea entries used in this study had smaller lesions than most of the wild species. Additional work is needed to determine if vegetative propagation enhanced disease susceptibility. In future studies, an ordinal rating ranking, such as the Florida scale for leaf spot (Chiteka et al., 1988), will added to accommodate informative qualitative characters such as wilting and death. Nonetheless, there were significant differences among the wild Arachis accessions in susceptibility to At. rolfsii as indicated by lesion length, mycelial growth, and AUDPCs. Within the wild species, A. microsperma PI 674407 and PI 666096 exhibited small lesions, mycelial growth, and AUDPC. Both accessions have previously shown resistance to Cylindrocladium black rot (Tallury et al., 2013). Arachis cardenasii PI 475994 and A. diogoi PI 468354 also had relatively small lesions. PI 475994 has high resistance to Meloidogyne javanica race 3 (Sharma et al., 2002), and PI 468354 is resistant to tomato spotted wilt virus (Lyerly et al., 2002; Wang et al., 2009). Since the three species have an A genome, facilitating introgression into cultivated peanut, these four accessions may merit additional evaluation as candidates for pre-breeding.