Table 1.
The range of dimensions and mean weight of seeds in the 1989 and 2014 samples captured on each of the U.S. Standard Testing Sieves. Sieve openings were given on the sieves. The 1989 samples were from the studies by Tschinkel [19,42]. Data for the 1989 seeds can be found in S1 Table.
Table 2.
Dates of seed burial and recovery, as well as source of seeds used. All dates except the last are 2015. Each of the six runs was composed of five replicates at four depths using four sizes of seeds (12+, 14, 18, 20).
Fig 1.
Size distribution of seeds and husks.
The size of seeds stored (A) and consumed (B) by P. badius. Size distribution of consumed seeds was determined from the chaff discarded on the nest disc by the ants, while that of stored seeds was determined directly by excavating nests and collecting their seeds. Stored seeds are mostly large, whereas consumed seeds are mostly small. Each line represents a colony.
Fig 2.
Representative seed and chaff samples.
(A) Representative samples of seeds taken from seed-carrying workers during colony relocations; (B and C) chaff taken from middens before or after move completion, respectively. The chaff represents seeds opened and used by the ants, whereas the moved seeds represent the stores in underground chambers.
Fig 3.
Comparison of the sizes of stored and used seeds.
The size distribution of seeds taken from seed-transporting workers during nest relocation, and seeds taken from the chaff. (n = 7 colonies). The chaff represents seeds opened and used by the ants, whereas the moved seeds represent the stores in underground chambers. Bars represent SE and whiskers 95% confidence intervals.
Fig 4.
The size distribution of seeds taken from returning foragers, and seeds taken from old and new chaff. The chaff represents seeds opened and used by the ants, and consists of seeds that are generally smaller than those brought in by foragers. (n = 6 colonies; representing 21 to 242 seeds per colony and 66 to 1011 chaff pieces per colony. Whiskers = 95% confidence intervals
Fig 5.
Seeds of four sizes marked with fluorescent ink.
A. Under visible light; B. Under ultraviolet light. Small = orange; medium = green; large = yellow; very large = blue.
Fig 6.
The location and disposition of four sizes of seeds in laboratory colonies.
Large and very large seeds were never opened and eaten, but were retrieved from the arena (top panel) and stored in the nest (middle panel). By contrast, small seeds, and to a lesser extent, medium seeds were rapidly husked, eaten and the husks discarded in the trash pile (bottom panel). Seed were retained on sieves as follows: small, No. 20; medium, No. 18, large, No. 14, very large, No. 12. Error bars = 95% confidence interval (n = 5 colonies)
Fig 7.
Final disposition of seeds in relation to seed size, as percent of the 30 offered seeds of each size.
A. and B. Large and very large seeds were retrieved from the arena and stored in the nest but were almost never opened and eaten, C. By contrast, small seeds, and to a lesser extent, medium seeds were rapidly husked, eaten and the husks discarded in the trash pile. N = 5 colonies. Seeds were retained on sieves as follows: small, No. 20; medium, No. 18, large, No. 14, very large, No. 12. Error bars = 95% confidence interval. ANOVA for seed size: in arena, F3,16 = 1.39; n.s.; unprocessed in nest, F3,16 = 6.57; p< 0.005; eaten, F3,16 = 100.5; p< 0.00001
Fig 8.
Seeds opened with and without majors present.
When majors were present in the nests, more seeds were opened, but the maximum size opened was the same as in nests lacking majors (n = 6 colonies).
Fig 9.
The appearance of marked seed husks in the middens of field colonies that had been offered marked germinating and non-germinating seeds.
Germinating seeds (blue symbols) were far more likely to appear in the middens than non-germinated (red symbols). Because husks were mostly only half of the seed coat, the number was divided by two to estimate whole seed-equivalents.
Fig 10.
(A) Germinating seeds; (B) dyed and marked germinating seeds; (C) dyed and marked germinating seeds under UV light. Note fluorescence of both the husk and the seed interior; and (D) drilled and dyed non-germinating seeds, broken open to reveal the blue-dyed tissue inside.
Fig 11.
The results of the dyed germinating and non-germinating seed experiment.
Dyed germinating seeds were fed to larvae in the laboratory, causing them to appear pink (A) and to fluoresce under UV light (B). From the midgut, the rhodamine dye is translocated into the rectum, appearing as pink and fluorescent dots at the posterior end of the larvae, like the light on a caboose.
Fig 12.
Results of the seed burial and laboratory germination experiments.
(A) Fraction of seeds germinating in the burial experiment, and (B) in the parallel laboratory temperature experiment. (A). Burial date and seed size together with their interactions accounted for 93% of the factor variance. Burial depth and its interactions accounted for only 6% of the factor variance, and are not shown in this graph. The similarity of the effects of laboratory temperature and burial date suggests that soil temperature is largely responsible for stimulating/inhibiting germination in the burial experiment. Mean soil temperature is shown as a thick, pale blue line in panel A with the scale on the right axis. (B). Incubation temperature and seed size together with their interactions accounted for 89% of the variance in germination rate. Run number and its interactions explained less than 11% of the factor variance and are not shown in this graph. Bars = 95% CI.
Fig 13.
Mean soil temperatures in the burial experiment.
Temperature were measured at the four burial depths during the burial periods. Although daily temperature variation differed greatly by depth, the mean temperatures were not significantly different by depth, but increased until August, and then decreased until the end of the year. The mean temperatures to which seeds were exposed at the different depths were thus similar.
Fig 14.
The daily progress of germination in the laboratory experiment, shown as percent germinating by elapsed day (n = 5).
The timing of germination differed for seeds of different sizes. The largest seeds (size 12) mostly germinated within less than 10 days, size 14 mostly germinated after a lag, but did not germinate at 32°C. Size 18 seeds showed a mixed pattern with the two higher temperatures inducing rapid germination, and the lower two slower and more prolonged germination. The smallest seeds (size 20) mostly germinated at low rates throughout the period at the two lower temperatures, but showed little germination at the two higher temperatures.
Fig 15.
When subterranean seed-storage chambers were divided with metals strips such that the ants had no access to one side, more germinating seeds were found on the access-denied side. Germinating seeds are circled in red. In this example, most are large.
Fig 16.
Composition by species of the 2014 seeds.
The size classes differed in the degree to which they were dominated by one or two species. Size 14 consisted of almost entirely Croton michauxi, which was also abundant in size 18 seeds. Size 18 and 20 had substantial species diversity, and size 12, although diverse, was strongly dominated by Diodia teres.
Fig 17.
The species of seeds recovered from 31 P. badius nests in 1989 [42].
The seeds were sifted by size, and the sieve on which they were retained is indicated under each image. The approximately 50 species of seeds will be identified in another publication. Several species varied in size and were retained on more than one sieve.
Fig 18.
Seed size distributions among chambers and nests.
Nine representative colonies out of the 31 in the 1989 samples are shown. Each line represents seed size (weight percent) distribution in a chamber, and each panel represents a colony. Size distributions differed strongly among colonies, but varied little among chambers within each colony. The colonies shown here were selected to illustrate the variety of patterns.
Fig 19.
A minor and major worker of P. badius shown at the same scale as a selection of the seeds stored in the nest.
In this image, the only seeds these ants can open are the three smallest sizes. Ant images modified from antweb.org. Photos by April Nobile: minor worker: URL: https://www.antweb.org/bigPicture.do?name=casent0104423&shot=p&number=1. Major worker: https://www.antweb.org/bigPicture.do?name=casent0103057&shot=p&number=1.
Fig 20.
Relative size of some of the seeds found in P. badius nests.
The seeds are shown in the order of their relative weight, relative to the smallest (lightest) seed, Polygonella gracilis. All seed species varied in weight, some substantially, so these relative weights and the order of weights is approximate.
Fig 21.
Comparison of the seed size distributions in 1989–90 with current distributions.
Small seeds are currently less abundant, and large ones relatively more abundant than in 1989, probably because of successional changes in the site. The ants thus are currently able to open a smaller proportion of the seeds than in 1989, and probably depend more heavily on seed germination. (Boxes = s.e., whiskers = 95% conf. intervals).
Fig 22.
The proportion of seed resources by seed size.
Weight percent of total in relation to seed size in the 31 colonies in the 1989 samples. Approximately 70% of seed biomass and 50% of seed numbers was contained in seeds that the ants cannot readily open. However, patterns differed greatly among colonies (Fig 18).