- Email: [email protected]
Rodent consumption and caching behaviour selects for specific seed traits
South African Journal of Botany 84 (2013) 83–87
Contents lists available at SciVerse ScienceDirect
South African Journal of Botany journal homepage: www.elsevier.com/locate/sajb
Rodent consumption and caching behaviour selects for specific seed traits U.D. Rusch a, J.J. Midgley b, B. Anderson a,⁎ a b
Department of Botany and Zoology, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa Department of Botany, University of Cape Town, University Private Bag, Rondebosch 7700, South Africa
a r t i c l e
i n f o
Article history: Received 23 June 2012 Received in revised form 17 September 2012 Accepted 20 September 2012 Available online 13 November 2012 Edited by JC Manning Keywords: Acomys subspinosus Cache Dispersal syndrome Fynbos Leucadendron sessile Scatter-hoarding Seed dispersal Seed selection
a b s t r a c t The seed characteristics selected for by scatter-hoarding rodents can have an impact on seed morphology, seedling establishment and ultimately on plant community structure. Using Leucadendron sessile (Proteaceae), it was recently discovered that rodents are seed dispersers in the fynbos biome of South Africa. However, little is known about the characteristics of rodent-dispersed seeds and the selective influence rodents have on seed morphology in this biome. We investigated the caching behaviour of rodents and asked whether variation in seed traits (size, hull thickness) influenced whether seeds were more likely to be consumed or cached. Rodents tended to disperse and bury, rather than consume, medium sized L. sessile seeds with medium hull thickness. In contrast, small or thin hulled seeds were preferentially eaten in situ and were seldom buried. Large seeds or seeds with thick hulls were often left untouched at depots. Our results suggest that rodents may impose stabilizing selective pressures on seed size and hull thickness, traits that may also have consequences for seedling mortality, dormancy, competitive interactions and the survival of fires. © 2012 SAAB. Published by Elsevier B.V. All rights reserved.
1. Introduction The propensity for organisms to converge upon similar traits under similar selective pressures is one of the hallmarks of natural selection. Perhaps some of the most striking and well documented examples of this are pollination guilds, in which groups of unrelated plants have adapted to the same pollination systems and share sets of recognizable floral traits (Van der Pijl, 1961). Also known as pollination syndromes, these traits are thought to be an adaptive response of the plant to a particular pollinator group (Van der Pijl, 1961). Similarly, seeds of different plants may converge upon a similar set of character traits as adaptations to the same dispersal vector, establishing a seed dispersal syndrome (Forget and Milleron, 1991). The fynbos biome of South Africa has two well-established seed dispersal syndromes associated with it: myrmecochory and serotiny. Myrmecochorous seeds have a fleshy elaiosome attached to the seed hull that serves as a food source for ants. The ants consume the elaiosome while the rest of the still viable seed is dispersed underground where it is protected from granivores and fires (e.g. Christian and Stanton, 2004). In contrast, serotinous seeds usually possess adaptations for wind dispersal and are only released from fireresistant cones after burning or death of the parent plant (Midgley et al., 2002). A third guild of fynbos plants possessing large-bodied, ⁎ Corresponding author. E-mail address: [email protected] (B. Anderson).
thick-hulled seeds, but without wings or elaiosomes has recently been identified as belonging to a putative rodent dispersal syndrome (Midgley and Anderson, 2005; Midgley et al., 2002). These seeds are often released by plants en masse during the summer months. This abundance of available seed over a short period of time satiates rodents and encourages caching behaviour (Forget, 1993; Vander Wall, 1990). Although not recorded in the fynbos, this release strategy is referred to as masting when plants accumulate seeds on a plant for several years before the seeds are dropped (Vander Wall, 1990). Seed selection behaviour for caching in rodents is tied to energetic and temporal efficiency — rodents are thought to cache seeds that are large and/or thick hulled (Brewer, 2001; Vander Wall, 1990). Small seeds are generally eaten on site as the low nutritional content does not warrant the energy expenditure of dispersal and burial (Jacobs, 1992; Jansen et al., 2004; Reichman and Fay, 1983; Theimer, 2003). Especially during masting events when rodents attempt to disperse and bury as many seeds as possible, small and thin hulled seeds are utilised as an immediate energy source since handling time for ingestion, e.g. cracking hulls is minimal (Forget, 1993; Vander Wall, 1990; Woodrey, 1990). Ingesting small or thin-hulled seeds allows the rodent to spend proportionally more time dispersing and burying larger, nutritionally valuable seeds. After preliminary observations of the Cape Spiny Mouse (Acomys subspinosus (Waterhouse, 1838)) burying seeds in captivity (Vlok, 1995), Midgley et al. (2002) confirmed that mice also disperse the large, thick-hulled seeds of the Proteiod plant, Leucadendron sessile
0254-6299/$ – see front matter © 2012 SAAB. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.sajb.2012.09.007
84
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
R.Br in the field. Midgley and Anderson (2005) subsequently found evidence suggesting that several other similarly large seeded plants belonging to the Leucadendron genus and Restionaceae family are also dispersed by rodents. They deduced that the Cape Spiny mouse was the most likely disperser in mountain fynbos areas while the hairy footed gerbil (Gerbillurus paeba (A. Smith, 1836)) dispersed seeds in sand plain fynbos. Here we investigate the possible role of seed morphology in determining seed selection for consumption versus burial in mountain fynbos. We manipulated seed size and hull thickness, two traits which may affect rodent handling efficiency (e.g. Theimer, 2003; Zang and Zang, 2008), and observed if rodent dispersal versus consumption behaviour changed with respective seed manipulations. Our expectation was that rodents would preferentially disperse and cache large, thick-hulled seeds but consume small, thin-hulled seeds in situ. 2. Materials and methods 2.1. Study site and species The study was conducted in April 2010 on the slopes of Sir Lowry's Pass near Somerset West, Western Cape, South Africa (200 m elevation). The field site comprised mostly undisturbed fynbos with an overgrown road transecting the site. Fynbos is a fire-dependant vegetation type, and species composition is affected by time elapsed since the last burn (Hoffman et al., 1987). The vegetation at the site was five years old. At the time of the study the site was dominated by L. sessile (suncone bushes), Protea repens and Willdenowia glomerata (Thunb.) H.P.Linder. L. sessile is a single-stemmed shrub growing to 2.0 meters tall. Flowering occurs in late winter (July– August) and seeds are dropped annually. The seed drop window extends from November to January with peak seed drop in December (early summer) (Manning, 2007). Directly after seed release, there are too many naturally released seeds in the environment to observe removal and caching of marked seeds. The best time to observe caching behaviour is when most of the seeds have been buried, but when rodents are still relatively satiated. The month of April is the best month to observe caching at Sir Lowrey's pass (Rusch, 2011). Three rodent species and one Macrosclided have been recorded at the research site: A. subspinosus, Rhabdomys pumilio (Spearman, 1784), Micaelamys namaquensis (A. Smith, 1834), and Elephantulus edwardii (A. Smith, 1839) (Midgley et al., 2002; UR, unpublished). M. namaquensis is a nocturnal omnivore and an opportunist granivore, whereas E. edwardii is a nocturnal insectivore not known to consume seeds (Chimimba and Bennett, 2005). Neither of the two species is known to disperse or bury seeds. R. pumilio is a diurnal omnivore and opportunist seed predator that does not cache seeds and dominates the diurnal rodent counts (Chimimba and Bennett, 2005; Rusch, 2011). The nocturnal A. subspinosus is the only known scatterhoarding rodent at the site, and dominated nocturnal population samples (Midgley et al., 2002; Rusch, 2011). A. subspinosus is a small (b 22 g), beige-grey rodent with white undersides, paws and dark, coarse hair running along the spine (Chimimba and Bennett, 2005). It is known to consume seeds and nectar (Fleming and Nicolson, 2002), but the details of its diet and seed handling behaviour are not well known (Midgley et al., 2002). 2.2. Seed preparation L. sessile seeds vary naturally in size and hull thickness. To determine the natural variation of seed traits for manipulation, we collected a single mature cone from 30 L. sessile plants. The seeds were shaken into a bag and 100 were removed. The seed hulls were cracked open and the thickest part of the hull was measured with callipers. The length, breadth and mass of the seed flesh was also measured with callipers. On average, seeds with their hulls on
weighed 0.48 g ± 0.4 (SE), with a minimum mass of 0.23 g and a maximum of 0.98 g. The average length of a de-hulled seeds was 4.2 mm ± 0.3 (SE), with a minimum and maximum of 2.7 mm and 7.9 mm respectively. The average breadth was 3.2 mm ± 0.2 (SE) with a minimum and maximum of 1.8 mm and 5.5 mm respectively. The average hull thickness was 2.0 mm ± 0.1 (SE) with a minimum and maximum thickness of 1.2 mm and 2.5 mm respectively. These baseline data were used to create standardized dimensions to manipulate experimental seeds. 2.2.1. Hull thickness As we were unable to determine hull thickness or flesh mass without damaging the seeds, we modified apparent hull thickness of seeds from a single plant. L. sessile seeds were coated with a layer of plumber's putty perforated to allow seed scent release. The wet putty was perforated ten times with a commercially available, medium sized paper clip. Plumber's putty is used commercially in association with drinking water and manufactured to be non-toxic and to dry odourless and tasteless. It has a comparable hardness and texture to natural seed hulls. We created three hull-thickness treatments based on our baseline data: 1. hull-thickness 50% thinner than mean hull-thickness (1 mm); 2. mean hull-thickness (2 mm); and 3. hulls 25% thicker than the mean (2.5 mm). In treatment 1 we removed 1 mm from the hull surface using sand paper (20 grit) by moving the seed over the full sheet of sand paper (30 cm) seven times on two sides of the seed. For treatment 2 we used unmodified L. sessile seeds; and for treatment 3 we applied a 0.5 mm thick layer of perforated putty over the seed. We used 100 seeds for each treatment, and fitted all seeds with a 30 cm fluorescent yellow tracking string. For all treatments, the string was attached to the seeds using a 2 mm by 2 mm piece of plumber's putty, controlling for the use of putty in the seed hull manipulations. Permanent pen was used to mark all tracking strings with stripes identifying each treatment and depot of origin, enabling us to record seeds that had been consumed. Tracking strings were used in previous studies (e.g. Midgley et al., 2002) and were not believed to have had an influence on seed selection by rodents. 2.2.2. Seed size Macadamia nuts (Macadamia integrifolia) were used to simulate different seed sizes by shaving down de-hulled, consumer grade nuts to the recorded smallest (2.7 mm× 1.8 mm), mean (4.2 mm× 3.2 mm) and largest dimensions (9.0 mm× 5.5 mm) found in naturally occurring L. sessile seeds. Macadamia and L. sessile seeds have a similar density (UR unpublished). All nuts were fitted with a standard 2 mm thick artificial hull of perforated plumber's putty. Macadamia nuts were used as they are similar in caloric value to L. sessile seeds (Rusch, 2011) and easy to shave down to the appropriate size. We prepared 100 seeds for each treatment and embedded a 30 cm fluorescent tracking string into all artificial hulls. Strings were marked with stripe-coding according to the manipulation. 2.3. Experimental design To investigate differences in mouse seed selection behaviour, we placed seeds with varying hull thickness and size into the field overnight and recorded seed fates the following morning (see below). Five transects were laid out perpendicular to the old overgrown road transecting the study site. Transect lines were 25 m apart, since A. subspinosus seldom disperses seeds more than 10 m (Midgley et al., 2002). Each transect line contained two seed depot sites (5 transects × 2 depot sites) with the first depot in the transect line approximately 5 m from the road and the second depot approximately 20 m from the first depot. All depot sites were placed under large, mature female L. sessile bushes. Depot sites were flat with slopes shallower than 15°.
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
2.3.2. Hull thickness trial Forty eight hours after the seed size trial with macadamia nuts, we returned to the field site and repeated the experiment described above with L. sessile seeds. We placed ten seeds of each hull thickness (thin, average and thick) at each depot. Thirty seeds per depot and a total of 300 seeds were used for this trial. Since L. sessile seeds were used two days after the macadamia nuts (seed size trial), the two seed species were never mixed at the depots. To ensure that the presence of plumber's putty had no effect on rodent behaviour, we compared the removal and ingestion rates of seeds with a natural seed coat and those with a plumber's putty coat. Removal and ingestion rates of average sized L. sessile seeds (natural seed coat) were compared with macadamia nuts of the same dimensions that were fitted with an artificial (plumber's putty) seed coat. Both natural and artificial seed coats were the same thickness. For this comparison, data from the seed size and hull-thickness experiments were used. 2.4. Statistical analysis Data sets were tested for homogeneity and normality, and the assumptions for ANOVA's with a Bonferroni post-hoc tests were met. We tested for differences in recovery percentages of each seed treatment within each seed fate group (not dispersed, dispersed and eaten). All percentage values were arc-sin root transformed. All analyses were conducted with SPSS 18.0 (SPSS 2010). A t-test was used for the seed coat comparisons, to examine the proportions of each seed fate between the naturally- and artificially-coated seeds. 3. Results Unmodified L. sessile seeds with average hull thickness were preferentially removed and buried by A. subspinosus (Fig. 1, ANOVA F10,13 = 16.94, p b 0.001). Thin-hulled L. sessile seeds had significantly higher in situ ingestion rates than average and thick-hulled L. sessile seeds (F10,13 = 13.34, p b 0.001). Thick-hulled seeds were most often left at the depot site compared to average and thin-hulled seeds (F10, 13 = 3.01, p b 0.05). Similar to the hull thickness trial, macadamia nuts that were modified to resemble average-sized/hulled L. sessile seeds, had significantly higher dispersal and burial rates than small and large seeds (Fig. 2, ANOVA, F10,13 = 8.93, p = 0.01). Small seeds had significantly higher in situ ingestion rates than average and large seeds (F10,13 = 5.55, p = 0.010) and large seeds were left at depot sites significantly
Percent (%) Seeds Recovered
Thin Hulled
100 90 80 70 60 50 40 30 20 10 0
a
Average Hulled
(37)
Thick Hulled b
c
(36)
(25)
c (19)
b (20) a
b
a
(2)
b
(10)
(4)
Not Dispersed
Dispersed & Buried
(4)
Eaten
Seed Fate Fig. 1. Fates of thin-hulled, average-hulled and thick-hulled L. sessile seeds after a 12 h exposure to A. subspinosus in the field. Thin-hulled seeds were eaten most often, average-hulled seeds cached and thick-hulled seeds were not dispersed. Percentage seeds recovered that share the same letter are not significant (p > 0.05), numbers in brackets indicate number of seeds recovered.
more often than small or average sized seeds (F10,13 = 3.24, p = 0.053). We found that there were no significant differences in the fates of medium macadamias and Leucadendron seeds fitted with different hulls (plumbers putty and natural hull respectively. The seeds were of similar size and did not differ in terms of whether they were dispersed or not (t37 = 2.46, p = 0.902), whether they were left at the depot untouched (t61 = 1.71, p = 0.122) or whether they were eaten (t22 = 3.03, p = 0.604). 4. Discussion 4.1. Seed dispersal Rodents clearly displayed preferences in terms of which seeds were dispersed or consumed. Seeds with average size and hull thickness dimensions (corresponding with mean values of natural L. sessile seeds) were preferentially dispersed and buried. Small or thin-hulled seeds were most often eaten in situ whereas large or thick-hulled seeds were frequently left at the depot sites. Selection against both extremes in terms of seed size and hull thickness suggests that rodents may be exerting stabilizing selective pressure on these traits. Since A. subspinosus is the only known scatter-hoarding rodent found at the site, it is most likely the primary cacher of seeds. Although other rodents may have contributed towards seed consumption, we suspect that most consumption was also due to A. subspinosus since the effects of the other important granivore (R. pumilio) were mitigated by running the experiments at night (R. pumilio is diurnal). It Small Seeds
Percent (%) Seeds Recovered
2.3.1. Seed size trial We placed ten manipulated macadamia seeds from each size class category (small seeds, average seeds and large seeds) making a total of 30 seeds per depot. Seeds were mixed and randomly dispersed over 1 m 2 under the respective depot sites. This procedure was repeated for all ten depot sites for a total of 300 seeds deposited. Deposition took place shortly before sunset and recovery took place shortly after sunrise the next morning to avoid the diurnally abundant R. pumilo, a species known to consume seeds without dispersing them (Midgley et al., 2002). Seed recovery was conducted by first checking a depot for non-dispersed/eaten seeds, before completing a full search in a 15 m radius around the depot site by walking in a spiral fashion away from the depot, recovering any dispersed, buried or eaten seeds. Seeds were presumed eaten when we found the hull cracked open and the seed inside missing. Usually the tracking string was still attached to part of the seed hull which allowed us to identify the type of seed that was ingested based on the stripe coding. If the seed was within 1 m of the depot site, intact and not buried, it was considered “not touched.” Seeds which the authors could not locate were recorded as “unknown fate.”
85
100 90 80 70 60 50 40 30 20 10 0
Average Seeds Large Seeds
a
b
(30)
b
(31)
(25) a
a (13) (18)
b b a (3)
Not Dispersed
a
(13)
(12)
(6)
Dispersed & Buried
Eaten
Seed Fate Fig. 2. The fates of small, average and large macadamia seeds after exposure to Acomys subspinosus for 12 h in the field. Small seeds were predominantly eaten, average sized seeds cached, and large seeds not dispersed. Percentage seeds recovered that share the same letter are not significantly different (p >0.05), numbers in brackets indicate the number of seeds recovered.
86
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
is possible that the nocturnal M. namaquensis was responsible for part of the seed consumption recorded in these experiments, however based on its low relative abundance at the field site (Rusch, 2011), this contribution is unlikely to be substantial. Birds and insects are not attracted to these seeds as they cannot access the rewards, therefore seed removal by these vectors was highly unlikely (Midgley et al., 2002). Plumber's putty was used on every experimental seed and was unlikely to have influenced the results. In the seed size experiment, the use of plumber's putty was completely controlled for, as all seeds were fitted with an identical 2 mm thick artificial hull made from putty. In the hull thickness experiments, we also used putty on all seeds, thus mitigating the “putty effect.” Although the different amounts of putty used in the hull thickness experiment may have influenced the results, we also showed that L. sessile seeds with natural hulls had similar fates to equally sized macadamia nuts fitted with artificial hulls of the same thickness. Furthermore the hull thickness experiment generated similar patterns to the seed size experiment where there could not have been a “putty effect.” These lines of evidence support the claim that the plumber's putty as an artificial seed hull had no effect on seed choice by rodents. However we acknowledge that a better way to have designed the hull thickness experiment would have been to completely de-hull all experimental seeds and fit all of them with complete artificial hulls of variable thickness. Other studies (e.g. Vander Wall, 2003) found that large seeds, which have a higher nutritional value than smaller seeds, should present the most energetically viable option for dispersal and burial in rodents (Kerley and Erasmus, 1991; Pulliam, 1974). However our results highlight the fact that handling, transport and burial costs should also be taken into consideration when looking at traitspecific seed selection. Large seeds require increased handling and transport time thereby increasing exposure risk to predators (Jacobs, 1992; Lima, 1985). Additionally, transport and handling are also energetically costly, suggesting that burial behaviour is a tradeoff: On the one hand seeds need to be large enough to warrant the energy expenditure of burial, but on the other hand, it is important that the costs of handling and exposure to predators should not exceed those benefits (Moore et al., 2007; Stapanian and Smith, 1984; Tamura et al., 1999). Given the relatively small body size of A. subspinosus (+/− 20 g) compared to other fynbos rodents such as M. namaquensis (+/− 60 g), the costs of handling larger seeds may rapidly outweigh the benefits. Jansen et al. (2002), Theimer (2003) as well as Munoz and Bonal (2008) all suggest a transport threshold where seeds above a certain percentage of the rodent's body weight are too heavy to carry efficiently, constraining or even preventing long-distance dispersal. The transport threshold seems to vary depending on the ecosystem and can be as little as 7% of the rodent's body mass (e.g. Jansen et al., 2002) or as much as 60% of rodent body mass (e.g. Munoz and Bonal, 2008). On average, unmanipulated L. sessile seeds weighed approximately 2.5% of A. subspinosus body mass with large seeds being close to the threshold observed by Jansen et al. (2002). Tamura and Hayashi (2008) also identified hull thickness and seed size as factors that may influence seed selection in mice. Seeds with thicker and/or harder hulls are often larger in size, making it increasingly difficult for smaller mice, like A. subspinosus, that have comparatively smaller masseter muscles and lower jaw bite-pressure (Zang and Zang, 2008), to gnaw into or crack these hulls. 4.2. Ecological significance Plants benefit from caching because it protects the seeds from the biotic environment (e.g. seed predators) as well as the abiotic environment (e.g. fires) (Vander Wall, 1990). Fires play a particularly important role in the fynbos since many seeds require fire to germinate
(Rebelo, 2001). Plants usually have adaptations for their seeds to survive fires, such as remaining in fire-resistant cones or exhibiting specific seed traits to encourage underground caching by animals. Due to their body size, rodents tend to disperse larger seeds than other vectors, such as wind or ants (Bond et al., 1999; Vander Wall, 2003). It is possible that large seeds may be advantageous if they enable germination from deep within the soils and therefore increase survival in particularly hot fires. Rodents bury their relatively large seeds to a maximum of 4 cm (Midgley et al., 2002). However, many nonrodent dispersed Leucospermums have much smaller seeds and are still able to germinate from similar depths (Bond et al., 1999). Despite similar germination depth ranges, large rodent-dispersed seeds may have an advantage over smaller seeds because their extra resources provide them with consistent and rapid growth rates after fire. This may allow them to survive the harsh, post-fire conditions better and out-compete seedlings establishing from smaller seed bodies. Additionally, rodent-dispersed seeds have thick hulls compared to the smaller ant or wind dispersed seeds, which may increase their resistance to heat as well as insect, fungi or bacterial attacks (Jansen and Forget, 2001). Holmes and Newton (2004) determined that Proteaceae seeds, some of which may be rodent-dispersed, have long-term persistence in the soil (half-life exceeding 2 years). Hull thickness and seed size may thus be important traits that affect mortality through fire-survival, surviving long phases of dormancy as well as competition. Acknowledgements We thank the Western Cape Nature Conservation Board, the South African Botanical Society and Fynbos Forum for funding through a Fynbos Forum Innovation Scholarship, as well as Danie Van Zyl and Lyndré Nel for countless hours of help in the field. We also thank John Manning for his very helpful editorial comments. References Bond, W.D., Honig, M., Maze, K.E., 1999. Seed size and seedling emergence: an allometric relationship and some ecological implications. Oceologia 120, 132–136. Brewer, S.W., 2001. Predation and dispersal of large and small seeds of a tropical palm. Oikos 92, 245–255. Chimimba, C.T., Bennett, N.C., 2005. Mammals of the Southern African Sub-Region. Cambridge University Press, Cape Town. Christian, C.E., Stanton, M.L., 2004. Cryptic consequences of a dispersal mutualism: seed burial, elaiosome removal and seed-bank dynamics. Ecology 85, 1101–1110. Fleming, P.A., Nicolson, S.W., 2002. Opportunistic breeding in the Cape Spiny Mouse (Acomys subspinosus). African Zoology 37, 101–105. Forget, P.M., 1993. Post-dispersal predation and scatterhoarding of Dipteryx panamensis (Papilionaceae) seeds by rodents in Panama. Oecologia 94, 255–261. Forget, P.M., Milleron, T., 1991. Evidence for secondary seed dispersal by rodents in Panama. Oecologia 87, 596–599. Hoffman, M.T., Moll, E.J., Boucher, C., 1987. Post-fire succession at Pella, a South African lowland fynbos site. South African Journal of Botany 53, 370–374. Holmes, P.M., Newton, R.J., 2004. Patterns of seed persistence in South African Fynbos. Plant Ecology 172, 143–158. Jacobs, L.F., 1992. The effect of handling time on the decision to cache by grey squirrels. Animal Behaviour 43, 522–524. Jansen, P.A., Forget, P.M., 2001. Scatter-hoarding rodents and tree regeneration in French Guiana. In: Bongers, F., Charles-Dominique, P., Forget, P.M., Thery, M. (Eds.), Nouragues: Dynamics and plant-animal interactions in a Neotropical rainforest. Kluwer Academic Publishers, Dordrecht, pp. 275–288. Jansen, P.A., Bartholomeus, M., Bongers, F., Elzinga, J.A., Den Ouden, J., van Wieren, S.E., 2002. The role of seed size in dispersal by a scatter-hoarding rodent. In: Levey, D.J., Sila, W.R., Galetti, M. (Eds.), Seed Dispersal and Frugivory: Ecology, Evolution and Conservation. CABI International, Brazil, pp. 209–225. Jansen, P.A., Bongers, F., Hemerik, L., 2004. Seed mass and mast seeding enhance dispersal by a neotropical scatter-hoarding rodent. Ecological Monographs 74, 569–589. Kerley, G.I.H., Erasmus, T., 1991. What do mice select for in seeds? Oecologia 86, 261–267. Lima, S.L., 1985. Maximizing feeding efficiency and minimizing time exposed to predators: a trade-off in the black capped chickadee. Oecologia 66, 60–67. Manning, J.C., 2007. Field Guide to Fynbos. Struik, Cape Town. Midgley, J.J., Anderson, B., 2005. Scatterhoarding in Mediterranean Shrublands of the SW Cape, South Africa. In: Forget, P.M. (Ed.), Seed Fate: Predation, Dispersal and Seedling Establishment. Ciba Publishing, pp. 197–204.
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87 Midgley, J.J., Anderson, B., Bok, A., Fleming, T., 2002. Scatter-hoarding of Cape Proteaceae nuts by rodents. Evolutionary Ecology Research 4, 623–626. Moore, J.E., McEuen, A.B., Swihart, R.K., Contreras, T.A., Steele, M.A., 2007. Determinants of seed removal distance by scatter-hoarding rodents in deciduous forests. Ecology 88, 2529–2540. Munoz, A., Bonal, R., 2008. Are you strong enough to carry that seed? Seed size/body size ratios influence seed choices by rodents. Animal Behaviour 76, 709–715. Pulliam, H.R., 1974. On the theory of optimal diets. The American Naturalist 108, 59–74. Rebelo, A., 2001. Proteas. A Field Guide to the Proteas of Southern Africa, 2nd ed. Fernwood Press, Cape Town. Reichman, O.J., Fay, P., 1983. Comparison of the diets of a caching and non-caching rodent. The American Naturalist 122, 576–581. Rusch, U.D., 2011.Scatter hoarding in Acomys subspinosus: the roles of seed traits, seasonality and cache retrieval. MSc Thesis, University of Stellenbosch, pp. 87. Stapanian, M.A., Smith, C.C., 1984. Density-dependent survival of scatterhoarded nuts: an experimental approach. Ecology 65, 1387–1396. Tamura, N., Hayashi, F., 2008. Geographic variation in walnut seed size correlates with hoarding behaviour of two rodent species. Ecological Research 23, 607–614.
87
Tamura, N., Hashimoto, Y., Hayashi, F., 1999. Optimal distances for squirrels to transport and hoard walnuts. Animal Behaviour 58, 635–642. Theimer, T.C., 2003. Intraspecific variation in seed size affects scatterhoarding behaviour of an Australian rain-forest rodent. Journal of Tropical Ecology 19, 95–98. Van der Pijl, L., 1961. Ecological aspects of flower evolution. II. Zoophilous flower classes. Evolution 15, 44–59. Vander Wall, S.B., 1990. Food Hoarding in Animals. University of Chicago Press, Chicago. Vander Wall, S.B., 2003. Effects of seed size of wind-dispersed pines (Pinus) on secondary seed dispersal and the caching behaviour of rodents. Oikos 100, 25–34. Vlok, J.H., 1995. Fynbos and rodents. Veld and Flora 81, 105–107. Woodrey, M.S., 1990. Economics of caching versus immediate consumption by WhiteBreasted Nuthatches: the effect of handling time. The Condor 92, 621–624. Zang, H., Zang, Z., 2008. Endocarp thickness affects seed removal speed by small rodents in a warm-temperate broad-leafed deciduous forest, China. Acta Oecologica 34, 283–285.
Contents lists available at SciVerse ScienceDirect
South African Journal of Botany journal homepage: www.elsevier.com/locate/sajb
Rodent consumption and caching behaviour selects for specific seed traits U.D. Rusch a, J.J. Midgley b, B. Anderson a,⁎ a b
Department of Botany and Zoology, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa Department of Botany, University of Cape Town, University Private Bag, Rondebosch 7700, South Africa
a r t i c l e
i n f o
Article history: Received 23 June 2012 Received in revised form 17 September 2012 Accepted 20 September 2012 Available online 13 November 2012 Edited by JC Manning Keywords: Acomys subspinosus Cache Dispersal syndrome Fynbos Leucadendron sessile Scatter-hoarding Seed dispersal Seed selection
a b s t r a c t The seed characteristics selected for by scatter-hoarding rodents can have an impact on seed morphology, seedling establishment and ultimately on plant community structure. Using Leucadendron sessile (Proteaceae), it was recently discovered that rodents are seed dispersers in the fynbos biome of South Africa. However, little is known about the characteristics of rodent-dispersed seeds and the selective influence rodents have on seed morphology in this biome. We investigated the caching behaviour of rodents and asked whether variation in seed traits (size, hull thickness) influenced whether seeds were more likely to be consumed or cached. Rodents tended to disperse and bury, rather than consume, medium sized L. sessile seeds with medium hull thickness. In contrast, small or thin hulled seeds were preferentially eaten in situ and were seldom buried. Large seeds or seeds with thick hulls were often left untouched at depots. Our results suggest that rodents may impose stabilizing selective pressures on seed size and hull thickness, traits that may also have consequences for seedling mortality, dormancy, competitive interactions and the survival of fires. © 2012 SAAB. Published by Elsevier B.V. All rights reserved.
1. Introduction The propensity for organisms to converge upon similar traits under similar selective pressures is one of the hallmarks of natural selection. Perhaps some of the most striking and well documented examples of this are pollination guilds, in which groups of unrelated plants have adapted to the same pollination systems and share sets of recognizable floral traits (Van der Pijl, 1961). Also known as pollination syndromes, these traits are thought to be an adaptive response of the plant to a particular pollinator group (Van der Pijl, 1961). Similarly, seeds of different plants may converge upon a similar set of character traits as adaptations to the same dispersal vector, establishing a seed dispersal syndrome (Forget and Milleron, 1991). The fynbos biome of South Africa has two well-established seed dispersal syndromes associated with it: myrmecochory and serotiny. Myrmecochorous seeds have a fleshy elaiosome attached to the seed hull that serves as a food source for ants. The ants consume the elaiosome while the rest of the still viable seed is dispersed underground where it is protected from granivores and fires (e.g. Christian and Stanton, 2004). In contrast, serotinous seeds usually possess adaptations for wind dispersal and are only released from fireresistant cones after burning or death of the parent plant (Midgley et al., 2002). A third guild of fynbos plants possessing large-bodied, ⁎ Corresponding author. E-mail address: [email protected] (B. Anderson).
thick-hulled seeds, but without wings or elaiosomes has recently been identified as belonging to a putative rodent dispersal syndrome (Midgley and Anderson, 2005; Midgley et al., 2002). These seeds are often released by plants en masse during the summer months. This abundance of available seed over a short period of time satiates rodents and encourages caching behaviour (Forget, 1993; Vander Wall, 1990). Although not recorded in the fynbos, this release strategy is referred to as masting when plants accumulate seeds on a plant for several years before the seeds are dropped (Vander Wall, 1990). Seed selection behaviour for caching in rodents is tied to energetic and temporal efficiency — rodents are thought to cache seeds that are large and/or thick hulled (Brewer, 2001; Vander Wall, 1990). Small seeds are generally eaten on site as the low nutritional content does not warrant the energy expenditure of dispersal and burial (Jacobs, 1992; Jansen et al., 2004; Reichman and Fay, 1983; Theimer, 2003). Especially during masting events when rodents attempt to disperse and bury as many seeds as possible, small and thin hulled seeds are utilised as an immediate energy source since handling time for ingestion, e.g. cracking hulls is minimal (Forget, 1993; Vander Wall, 1990; Woodrey, 1990). Ingesting small or thin-hulled seeds allows the rodent to spend proportionally more time dispersing and burying larger, nutritionally valuable seeds. After preliminary observations of the Cape Spiny Mouse (Acomys subspinosus (Waterhouse, 1838)) burying seeds in captivity (Vlok, 1995), Midgley et al. (2002) confirmed that mice also disperse the large, thick-hulled seeds of the Proteiod plant, Leucadendron sessile
0254-6299/$ – see front matter © 2012 SAAB. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.sajb.2012.09.007
84
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
R.Br in the field. Midgley and Anderson (2005) subsequently found evidence suggesting that several other similarly large seeded plants belonging to the Leucadendron genus and Restionaceae family are also dispersed by rodents. They deduced that the Cape Spiny mouse was the most likely disperser in mountain fynbos areas while the hairy footed gerbil (Gerbillurus paeba (A. Smith, 1836)) dispersed seeds in sand plain fynbos. Here we investigate the possible role of seed morphology in determining seed selection for consumption versus burial in mountain fynbos. We manipulated seed size and hull thickness, two traits which may affect rodent handling efficiency (e.g. Theimer, 2003; Zang and Zang, 2008), and observed if rodent dispersal versus consumption behaviour changed with respective seed manipulations. Our expectation was that rodents would preferentially disperse and cache large, thick-hulled seeds but consume small, thin-hulled seeds in situ. 2. Materials and methods 2.1. Study site and species The study was conducted in April 2010 on the slopes of Sir Lowry's Pass near Somerset West, Western Cape, South Africa (200 m elevation). The field site comprised mostly undisturbed fynbos with an overgrown road transecting the site. Fynbos is a fire-dependant vegetation type, and species composition is affected by time elapsed since the last burn (Hoffman et al., 1987). The vegetation at the site was five years old. At the time of the study the site was dominated by L. sessile (suncone bushes), Protea repens and Willdenowia glomerata (Thunb.) H.P.Linder. L. sessile is a single-stemmed shrub growing to 2.0 meters tall. Flowering occurs in late winter (July– August) and seeds are dropped annually. The seed drop window extends from November to January with peak seed drop in December (early summer) (Manning, 2007). Directly after seed release, there are too many naturally released seeds in the environment to observe removal and caching of marked seeds. The best time to observe caching behaviour is when most of the seeds have been buried, but when rodents are still relatively satiated. The month of April is the best month to observe caching at Sir Lowrey's pass (Rusch, 2011). Three rodent species and one Macrosclided have been recorded at the research site: A. subspinosus, Rhabdomys pumilio (Spearman, 1784), Micaelamys namaquensis (A. Smith, 1834), and Elephantulus edwardii (A. Smith, 1839) (Midgley et al., 2002; UR, unpublished). M. namaquensis is a nocturnal omnivore and an opportunist granivore, whereas E. edwardii is a nocturnal insectivore not known to consume seeds (Chimimba and Bennett, 2005). Neither of the two species is known to disperse or bury seeds. R. pumilio is a diurnal omnivore and opportunist seed predator that does not cache seeds and dominates the diurnal rodent counts (Chimimba and Bennett, 2005; Rusch, 2011). The nocturnal A. subspinosus is the only known scatterhoarding rodent at the site, and dominated nocturnal population samples (Midgley et al., 2002; Rusch, 2011). A. subspinosus is a small (b 22 g), beige-grey rodent with white undersides, paws and dark, coarse hair running along the spine (Chimimba and Bennett, 2005). It is known to consume seeds and nectar (Fleming and Nicolson, 2002), but the details of its diet and seed handling behaviour are not well known (Midgley et al., 2002). 2.2. Seed preparation L. sessile seeds vary naturally in size and hull thickness. To determine the natural variation of seed traits for manipulation, we collected a single mature cone from 30 L. sessile plants. The seeds were shaken into a bag and 100 were removed. The seed hulls were cracked open and the thickest part of the hull was measured with callipers. The length, breadth and mass of the seed flesh was also measured with callipers. On average, seeds with their hulls on
weighed 0.48 g ± 0.4 (SE), with a minimum mass of 0.23 g and a maximum of 0.98 g. The average length of a de-hulled seeds was 4.2 mm ± 0.3 (SE), with a minimum and maximum of 2.7 mm and 7.9 mm respectively. The average breadth was 3.2 mm ± 0.2 (SE) with a minimum and maximum of 1.8 mm and 5.5 mm respectively. The average hull thickness was 2.0 mm ± 0.1 (SE) with a minimum and maximum thickness of 1.2 mm and 2.5 mm respectively. These baseline data were used to create standardized dimensions to manipulate experimental seeds. 2.2.1. Hull thickness As we were unable to determine hull thickness or flesh mass without damaging the seeds, we modified apparent hull thickness of seeds from a single plant. L. sessile seeds were coated with a layer of plumber's putty perforated to allow seed scent release. The wet putty was perforated ten times with a commercially available, medium sized paper clip. Plumber's putty is used commercially in association with drinking water and manufactured to be non-toxic and to dry odourless and tasteless. It has a comparable hardness and texture to natural seed hulls. We created three hull-thickness treatments based on our baseline data: 1. hull-thickness 50% thinner than mean hull-thickness (1 mm); 2. mean hull-thickness (2 mm); and 3. hulls 25% thicker than the mean (2.5 mm). In treatment 1 we removed 1 mm from the hull surface using sand paper (20 grit) by moving the seed over the full sheet of sand paper (30 cm) seven times on two sides of the seed. For treatment 2 we used unmodified L. sessile seeds; and for treatment 3 we applied a 0.5 mm thick layer of perforated putty over the seed. We used 100 seeds for each treatment, and fitted all seeds with a 30 cm fluorescent yellow tracking string. For all treatments, the string was attached to the seeds using a 2 mm by 2 mm piece of plumber's putty, controlling for the use of putty in the seed hull manipulations. Permanent pen was used to mark all tracking strings with stripes identifying each treatment and depot of origin, enabling us to record seeds that had been consumed. Tracking strings were used in previous studies (e.g. Midgley et al., 2002) and were not believed to have had an influence on seed selection by rodents. 2.2.2. Seed size Macadamia nuts (Macadamia integrifolia) were used to simulate different seed sizes by shaving down de-hulled, consumer grade nuts to the recorded smallest (2.7 mm× 1.8 mm), mean (4.2 mm× 3.2 mm) and largest dimensions (9.0 mm× 5.5 mm) found in naturally occurring L. sessile seeds. Macadamia and L. sessile seeds have a similar density (UR unpublished). All nuts were fitted with a standard 2 mm thick artificial hull of perforated plumber's putty. Macadamia nuts were used as they are similar in caloric value to L. sessile seeds (Rusch, 2011) and easy to shave down to the appropriate size. We prepared 100 seeds for each treatment and embedded a 30 cm fluorescent tracking string into all artificial hulls. Strings were marked with stripe-coding according to the manipulation. 2.3. Experimental design To investigate differences in mouse seed selection behaviour, we placed seeds with varying hull thickness and size into the field overnight and recorded seed fates the following morning (see below). Five transects were laid out perpendicular to the old overgrown road transecting the study site. Transect lines were 25 m apart, since A. subspinosus seldom disperses seeds more than 10 m (Midgley et al., 2002). Each transect line contained two seed depot sites (5 transects × 2 depot sites) with the first depot in the transect line approximately 5 m from the road and the second depot approximately 20 m from the first depot. All depot sites were placed under large, mature female L. sessile bushes. Depot sites were flat with slopes shallower than 15°.
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
2.3.2. Hull thickness trial Forty eight hours after the seed size trial with macadamia nuts, we returned to the field site and repeated the experiment described above with L. sessile seeds. We placed ten seeds of each hull thickness (thin, average and thick) at each depot. Thirty seeds per depot and a total of 300 seeds were used for this trial. Since L. sessile seeds were used two days after the macadamia nuts (seed size trial), the two seed species were never mixed at the depots. To ensure that the presence of plumber's putty had no effect on rodent behaviour, we compared the removal and ingestion rates of seeds with a natural seed coat and those with a plumber's putty coat. Removal and ingestion rates of average sized L. sessile seeds (natural seed coat) were compared with macadamia nuts of the same dimensions that were fitted with an artificial (plumber's putty) seed coat. Both natural and artificial seed coats were the same thickness. For this comparison, data from the seed size and hull-thickness experiments were used. 2.4. Statistical analysis Data sets were tested for homogeneity and normality, and the assumptions for ANOVA's with a Bonferroni post-hoc tests were met. We tested for differences in recovery percentages of each seed treatment within each seed fate group (not dispersed, dispersed and eaten). All percentage values were arc-sin root transformed. All analyses were conducted with SPSS 18.0 (SPSS 2010). A t-test was used for the seed coat comparisons, to examine the proportions of each seed fate between the naturally- and artificially-coated seeds. 3. Results Unmodified L. sessile seeds with average hull thickness were preferentially removed and buried by A. subspinosus (Fig. 1, ANOVA F10,13 = 16.94, p b 0.001). Thin-hulled L. sessile seeds had significantly higher in situ ingestion rates than average and thick-hulled L. sessile seeds (F10,13 = 13.34, p b 0.001). Thick-hulled seeds were most often left at the depot site compared to average and thin-hulled seeds (F10, 13 = 3.01, p b 0.05). Similar to the hull thickness trial, macadamia nuts that were modified to resemble average-sized/hulled L. sessile seeds, had significantly higher dispersal and burial rates than small and large seeds (Fig. 2, ANOVA, F10,13 = 8.93, p = 0.01). Small seeds had significantly higher in situ ingestion rates than average and large seeds (F10,13 = 5.55, p = 0.010) and large seeds were left at depot sites significantly
Percent (%) Seeds Recovered
Thin Hulled
100 90 80 70 60 50 40 30 20 10 0
a
Average Hulled
(37)
Thick Hulled b
c
(36)
(25)
c (19)
b (20) a
b
a
(2)
b
(10)
(4)
Not Dispersed
Dispersed & Buried
(4)
Eaten
Seed Fate Fig. 1. Fates of thin-hulled, average-hulled and thick-hulled L. sessile seeds after a 12 h exposure to A. subspinosus in the field. Thin-hulled seeds were eaten most often, average-hulled seeds cached and thick-hulled seeds were not dispersed. Percentage seeds recovered that share the same letter are not significant (p > 0.05), numbers in brackets indicate number of seeds recovered.
more often than small or average sized seeds (F10,13 = 3.24, p = 0.053). We found that there were no significant differences in the fates of medium macadamias and Leucadendron seeds fitted with different hulls (plumbers putty and natural hull respectively. The seeds were of similar size and did not differ in terms of whether they were dispersed or not (t37 = 2.46, p = 0.902), whether they were left at the depot untouched (t61 = 1.71, p = 0.122) or whether they were eaten (t22 = 3.03, p = 0.604). 4. Discussion 4.1. Seed dispersal Rodents clearly displayed preferences in terms of which seeds were dispersed or consumed. Seeds with average size and hull thickness dimensions (corresponding with mean values of natural L. sessile seeds) were preferentially dispersed and buried. Small or thin-hulled seeds were most often eaten in situ whereas large or thick-hulled seeds were frequently left at the depot sites. Selection against both extremes in terms of seed size and hull thickness suggests that rodents may be exerting stabilizing selective pressure on these traits. Since A. subspinosus is the only known scatter-hoarding rodent found at the site, it is most likely the primary cacher of seeds. Although other rodents may have contributed towards seed consumption, we suspect that most consumption was also due to A. subspinosus since the effects of the other important granivore (R. pumilio) were mitigated by running the experiments at night (R. pumilio is diurnal). It Small Seeds
Percent (%) Seeds Recovered
2.3.1. Seed size trial We placed ten manipulated macadamia seeds from each size class category (small seeds, average seeds and large seeds) making a total of 30 seeds per depot. Seeds were mixed and randomly dispersed over 1 m 2 under the respective depot sites. This procedure was repeated for all ten depot sites for a total of 300 seeds deposited. Deposition took place shortly before sunset and recovery took place shortly after sunrise the next morning to avoid the diurnally abundant R. pumilo, a species known to consume seeds without dispersing them (Midgley et al., 2002). Seed recovery was conducted by first checking a depot for non-dispersed/eaten seeds, before completing a full search in a 15 m radius around the depot site by walking in a spiral fashion away from the depot, recovering any dispersed, buried or eaten seeds. Seeds were presumed eaten when we found the hull cracked open and the seed inside missing. Usually the tracking string was still attached to part of the seed hull which allowed us to identify the type of seed that was ingested based on the stripe coding. If the seed was within 1 m of the depot site, intact and not buried, it was considered “not touched.” Seeds which the authors could not locate were recorded as “unknown fate.”
85
100 90 80 70 60 50 40 30 20 10 0
Average Seeds Large Seeds
a
b
(30)
b
(31)
(25) a
a (13) (18)
b b a (3)
Not Dispersed
a
(13)
(12)
(6)
Dispersed & Buried
Eaten
Seed Fate Fig. 2. The fates of small, average and large macadamia seeds after exposure to Acomys subspinosus for 12 h in the field. Small seeds were predominantly eaten, average sized seeds cached, and large seeds not dispersed. Percentage seeds recovered that share the same letter are not significantly different (p >0.05), numbers in brackets indicate the number of seeds recovered.
86
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87
is possible that the nocturnal M. namaquensis was responsible for part of the seed consumption recorded in these experiments, however based on its low relative abundance at the field site (Rusch, 2011), this contribution is unlikely to be substantial. Birds and insects are not attracted to these seeds as they cannot access the rewards, therefore seed removal by these vectors was highly unlikely (Midgley et al., 2002). Plumber's putty was used on every experimental seed and was unlikely to have influenced the results. In the seed size experiment, the use of plumber's putty was completely controlled for, as all seeds were fitted with an identical 2 mm thick artificial hull made from putty. In the hull thickness experiments, we also used putty on all seeds, thus mitigating the “putty effect.” Although the different amounts of putty used in the hull thickness experiment may have influenced the results, we also showed that L. sessile seeds with natural hulls had similar fates to equally sized macadamia nuts fitted with artificial hulls of the same thickness. Furthermore the hull thickness experiment generated similar patterns to the seed size experiment where there could not have been a “putty effect.” These lines of evidence support the claim that the plumber's putty as an artificial seed hull had no effect on seed choice by rodents. However we acknowledge that a better way to have designed the hull thickness experiment would have been to completely de-hull all experimental seeds and fit all of them with complete artificial hulls of variable thickness. Other studies (e.g. Vander Wall, 2003) found that large seeds, which have a higher nutritional value than smaller seeds, should present the most energetically viable option for dispersal and burial in rodents (Kerley and Erasmus, 1991; Pulliam, 1974). However our results highlight the fact that handling, transport and burial costs should also be taken into consideration when looking at traitspecific seed selection. Large seeds require increased handling and transport time thereby increasing exposure risk to predators (Jacobs, 1992; Lima, 1985). Additionally, transport and handling are also energetically costly, suggesting that burial behaviour is a tradeoff: On the one hand seeds need to be large enough to warrant the energy expenditure of burial, but on the other hand, it is important that the costs of handling and exposure to predators should not exceed those benefits (Moore et al., 2007; Stapanian and Smith, 1984; Tamura et al., 1999). Given the relatively small body size of A. subspinosus (+/− 20 g) compared to other fynbos rodents such as M. namaquensis (+/− 60 g), the costs of handling larger seeds may rapidly outweigh the benefits. Jansen et al. (2002), Theimer (2003) as well as Munoz and Bonal (2008) all suggest a transport threshold where seeds above a certain percentage of the rodent's body weight are too heavy to carry efficiently, constraining or even preventing long-distance dispersal. The transport threshold seems to vary depending on the ecosystem and can be as little as 7% of the rodent's body mass (e.g. Jansen et al., 2002) or as much as 60% of rodent body mass (e.g. Munoz and Bonal, 2008). On average, unmanipulated L. sessile seeds weighed approximately 2.5% of A. subspinosus body mass with large seeds being close to the threshold observed by Jansen et al. (2002). Tamura and Hayashi (2008) also identified hull thickness and seed size as factors that may influence seed selection in mice. Seeds with thicker and/or harder hulls are often larger in size, making it increasingly difficult for smaller mice, like A. subspinosus, that have comparatively smaller masseter muscles and lower jaw bite-pressure (Zang and Zang, 2008), to gnaw into or crack these hulls. 4.2. Ecological significance Plants benefit from caching because it protects the seeds from the biotic environment (e.g. seed predators) as well as the abiotic environment (e.g. fires) (Vander Wall, 1990). Fires play a particularly important role in the fynbos since many seeds require fire to germinate
(Rebelo, 2001). Plants usually have adaptations for their seeds to survive fires, such as remaining in fire-resistant cones or exhibiting specific seed traits to encourage underground caching by animals. Due to their body size, rodents tend to disperse larger seeds than other vectors, such as wind or ants (Bond et al., 1999; Vander Wall, 2003). It is possible that large seeds may be advantageous if they enable germination from deep within the soils and therefore increase survival in particularly hot fires. Rodents bury their relatively large seeds to a maximum of 4 cm (Midgley et al., 2002). However, many nonrodent dispersed Leucospermums have much smaller seeds and are still able to germinate from similar depths (Bond et al., 1999). Despite similar germination depth ranges, large rodent-dispersed seeds may have an advantage over smaller seeds because their extra resources provide them with consistent and rapid growth rates after fire. This may allow them to survive the harsh, post-fire conditions better and out-compete seedlings establishing from smaller seed bodies. Additionally, rodent-dispersed seeds have thick hulls compared to the smaller ant or wind dispersed seeds, which may increase their resistance to heat as well as insect, fungi or bacterial attacks (Jansen and Forget, 2001). Holmes and Newton (2004) determined that Proteaceae seeds, some of which may be rodent-dispersed, have long-term persistence in the soil (half-life exceeding 2 years). Hull thickness and seed size may thus be important traits that affect mortality through fire-survival, surviving long phases of dormancy as well as competition. Acknowledgements We thank the Western Cape Nature Conservation Board, the South African Botanical Society and Fynbos Forum for funding through a Fynbos Forum Innovation Scholarship, as well as Danie Van Zyl and Lyndré Nel for countless hours of help in the field. We also thank John Manning for his very helpful editorial comments. References Bond, W.D., Honig, M., Maze, K.E., 1999. Seed size and seedling emergence: an allometric relationship and some ecological implications. Oceologia 120, 132–136. Brewer, S.W., 2001. Predation and dispersal of large and small seeds of a tropical palm. Oikos 92, 245–255. Chimimba, C.T., Bennett, N.C., 2005. Mammals of the Southern African Sub-Region. Cambridge University Press, Cape Town. Christian, C.E., Stanton, M.L., 2004. Cryptic consequences of a dispersal mutualism: seed burial, elaiosome removal and seed-bank dynamics. Ecology 85, 1101–1110. Fleming, P.A., Nicolson, S.W., 2002. Opportunistic breeding in the Cape Spiny Mouse (Acomys subspinosus). African Zoology 37, 101–105. Forget, P.M., 1993. Post-dispersal predation and scatterhoarding of Dipteryx panamensis (Papilionaceae) seeds by rodents in Panama. Oecologia 94, 255–261. Forget, P.M., Milleron, T., 1991. Evidence for secondary seed dispersal by rodents in Panama. Oecologia 87, 596–599. Hoffman, M.T., Moll, E.J., Boucher, C., 1987. Post-fire succession at Pella, a South African lowland fynbos site. South African Journal of Botany 53, 370–374. Holmes, P.M., Newton, R.J., 2004. Patterns of seed persistence in South African Fynbos. Plant Ecology 172, 143–158. Jacobs, L.F., 1992. The effect of handling time on the decision to cache by grey squirrels. Animal Behaviour 43, 522–524. Jansen, P.A., Forget, P.M., 2001. Scatter-hoarding rodents and tree regeneration in French Guiana. In: Bongers, F., Charles-Dominique, P., Forget, P.M., Thery, M. (Eds.), Nouragues: Dynamics and plant-animal interactions in a Neotropical rainforest. Kluwer Academic Publishers, Dordrecht, pp. 275–288. Jansen, P.A., Bartholomeus, M., Bongers, F., Elzinga, J.A., Den Ouden, J., van Wieren, S.E., 2002. The role of seed size in dispersal by a scatter-hoarding rodent. In: Levey, D.J., Sila, W.R., Galetti, M. (Eds.), Seed Dispersal and Frugivory: Ecology, Evolution and Conservation. CABI International, Brazil, pp. 209–225. Jansen, P.A., Bongers, F., Hemerik, L., 2004. Seed mass and mast seeding enhance dispersal by a neotropical scatter-hoarding rodent. Ecological Monographs 74, 569–589. Kerley, G.I.H., Erasmus, T., 1991. What do mice select for in seeds? Oecologia 86, 261–267. Lima, S.L., 1985. Maximizing feeding efficiency and minimizing time exposed to predators: a trade-off in the black capped chickadee. Oecologia 66, 60–67. Manning, J.C., 2007. Field Guide to Fynbos. Struik, Cape Town. Midgley, J.J., Anderson, B., 2005. Scatterhoarding in Mediterranean Shrublands of the SW Cape, South Africa. In: Forget, P.M. (Ed.), Seed Fate: Predation, Dispersal and Seedling Establishment. Ciba Publishing, pp. 197–204.
U.D. Rusch et al. / South African Journal of Botany 84 (2013) 83–87 Midgley, J.J., Anderson, B., Bok, A., Fleming, T., 2002. Scatter-hoarding of Cape Proteaceae nuts by rodents. Evolutionary Ecology Research 4, 623–626. Moore, J.E., McEuen, A.B., Swihart, R.K., Contreras, T.A., Steele, M.A., 2007. Determinants of seed removal distance by scatter-hoarding rodents in deciduous forests. Ecology 88, 2529–2540. Munoz, A., Bonal, R., 2008. Are you strong enough to carry that seed? Seed size/body size ratios influence seed choices by rodents. Animal Behaviour 76, 709–715. Pulliam, H.R., 1974. On the theory of optimal diets. The American Naturalist 108, 59–74. Rebelo, A., 2001. Proteas. A Field Guide to the Proteas of Southern Africa, 2nd ed. Fernwood Press, Cape Town. Reichman, O.J., Fay, P., 1983. Comparison of the diets of a caching and non-caching rodent. The American Naturalist 122, 576–581. Rusch, U.D., 2011.Scatter hoarding in Acomys subspinosus: the roles of seed traits, seasonality and cache retrieval. MSc Thesis, University of Stellenbosch, pp. 87. Stapanian, M.A., Smith, C.C., 1984. Density-dependent survival of scatterhoarded nuts: an experimental approach. Ecology 65, 1387–1396. Tamura, N., Hayashi, F., 2008. Geographic variation in walnut seed size correlates with hoarding behaviour of two rodent species. Ecological Research 23, 607–614.
87
Tamura, N., Hashimoto, Y., Hayashi, F., 1999. Optimal distances for squirrels to transport and hoard walnuts. Animal Behaviour 58, 635–642. Theimer, T.C., 2003. Intraspecific variation in seed size affects scatterhoarding behaviour of an Australian rain-forest rodent. Journal of Tropical Ecology 19, 95–98. Van der Pijl, L., 1961. Ecological aspects of flower evolution. II. Zoophilous flower classes. Evolution 15, 44–59. Vander Wall, S.B., 1990. Food Hoarding in Animals. University of Chicago Press, Chicago. Vander Wall, S.B., 2003. Effects of seed size of wind-dispersed pines (Pinus) on secondary seed dispersal and the caching behaviour of rodents. Oikos 100, 25–34. Vlok, J.H., 1995. Fynbos and rodents. Veld and Flora 81, 105–107. Woodrey, M.S., 1990. Economics of caching versus immediate consumption by WhiteBreasted Nuthatches: the effect of handling time. The Condor 92, 621–624. Zang, H., Zang, Z., 2008. Endocarp thickness affects seed removal speed by small rodents in a warm-temperate broad-leafed deciduous forest, China. Acta Oecologica 34, 283–285.