Protea Atlas Logo
  Overview of Project
  Project Staff
  Checking, Illustrations
  Upcoming Activities
  Id and  Species Lists
  Protea Information
  Protea Gallery
  Growing Proteas
  Interim Dist. Maps
  Afrikaanse Inligting


Scatter-Hoarding of Conebush Nuts by the Cape Spiny Mouse

Protea Atlas Logoby Jeremy Midgley, UCT (reworked from abstracts and MS by TR)

Cape members of the Protea Family generally have seeds that are either ant-dispersed (myrmecochorous), such as pincushions, or are stored away in closed cones (such as many proteas). In both cases, these seeds are considered to be safe from predation by seed-eating rodents. Studies in the past have shown that removal of seeds by rodents is very high. This has been viewed as the major driving force behind the evolution of these unusual seed dispersal and storage mechanisms. And yet, curiously, a third strategy exists in a group of Conebushes. These species merely drop their seeds onto the soil surface en masse each year. Why don’t the rodents just eat them all up?

I’ve long wondered about how seeds of these species survive. Might they actually be rodent dispersed? William Bond and Johan Breytenbach noted that many captive small mammals, especially the Cape Spiny Mouse, are good at locating buried seeds and Jan Vlok observed a caged Cape Spiny Mouse burying sunflower seeds. Tony Rebelo and John Rourke proposed in a review of seed dispersal and germination strategies in proteas in 1985, that the Conebushes with large, rounded nuts might be dispersed by rodents. However, to date there has been no field data to support these ideas. So, together with Bruce Anderson and Adele Bok, I eventually decided to see what "game" these species are playing.

We chose Sir Lowry’s Pass as our study area as it has a big population of Ld sessile, which has large nuts. The veld also contains burned and unburned areas. Our problem was how to keep track of our seeds – they need to be relocatable. So we superglued about 25 cm of brightly coloured fly-fishing line onto hundreds of Ld sessile nuts and placed them out in the veld. To our amazement, the following day most (more than 90%) had been removed. Most had been neatly buried about 10-20 mm deep. The seeds were very carefully buried, with hardly any signs of digging or soil disturbance - they looked almost as if they had merely been pushed below the soil surface. Occasionally two seeds were buried together, but most were alone. This is evidence that they really are scatter-hoarded (an alternative is larder hoarding, where seeds are stored in a central burrow – but this would not be a good dispersal strategy for Conebushes). Less than 1% of seeds recovered were eaten (but we only found half of them, so this could be quite high, but the vegetation was dense so we may have missed some, or the string may have been bitten off making them untraceable). More than three-quarters of these seeds were moved more than 2 m.

We could not find all the tagged seeds we put out, suggesting some are moved a long way sideways or a long way downwards. We tried to locate these missing seeds with another trial by gluing on 2m long tails. Again removal rates were high but again we could not locate all the tagged seeds. Some were taken deep into burrows and these had generally been eaten. We wondered how many seeds could be taken from a single point, so we placed piles out of up to 50 seeds. Next morning all had been taken from some piles. This suggests a serious hoarding behaviour.

To identify the rodent we used live-traps, baited with seeds, and caught many Acomys subspinosus Cape Spiny Mouse individuals, some of which were pregnant. These were released unharmed. These results suggest that the Cape Spiny Mouse may time birth to the bonanza time when seeds are released. In the laboratory we also observed them to be adept at finding buried Silvertree nuts. They don’t seem to make tunnels, but they are clearly keen diggers.

We looked at seedlings coming up from the recent fire and found a mean seed burial depth of 21 mm. Fewer than 2% of seedlings emerged from deeper than 40 mm suggesting that very few seeds are larder-hoarded in burrows. At Sir Lowry’s Pass we found female bushes produce about 15.9 cones, with about 8.4 seeds per cone. Thus each female produces 135 seeds per year at maturity, which contrasted with the 4.9 seedlings found per burned female skeleton. This suggests (unless seeds remain in the seed bank, and surely some must) a maximum predation rate of 96.4%.

This finding raises many interesting questions. Are nut-fruited species constrained in their distribution patterns by that of Acomys? How important are nuts to Acomys? How good is Acomys at finding the nuts it had previously buried? Why be ant-dispersed and how many ant-buried seeds are found by Acomys?

An even more interesting question is why do our Mice scatter rather than larder horde? It may be that being a small mouse, larders may be raided by bigger rodents, making scatterred hordes a safer bet. Curiously, our ants also behave strangely, in that they conveniently eat the elaiosomes off the fruit underground, instead of just eating them off above ground (as do alien invader ants). Might this be to get the seeds out of the way of the mice, which would happily steal them if ants hung around trying to take off the elaiosome above ground.

Nut-fruits of unknown dispersal also occur in other families, (e.g. Restionaceae: Ceratocaryum argenteum). It is thus likely that scatter-hoarding is fairly widespread. The Cape Spiny Mouse also buried tagged serotinous seeds of Ld laureolum and myrmecochorous seeds of Ls conocarpodendron, and even of sunflower nuts.

Clearly, the "avoiding-rodent-access" paradigm must be re-evaluated as the main reason for the evolution of serotiny and myrmecochory. Rodent seed-burial is similar in depth to that of ant-burial and post-fire regeneration rates are comparable to that of serotinous species. Germination cues are identical to that of ant-dispersed species. The big question though is why a plant in a putatively nutrient-poor environment can afford to feed over 90% of its seeds to mice? Why do proteas and pincushions with serotiny and myrmecochory – which avoid such predation - not eliminate species with such wasteful strategies in the numbers game of regeneration.

Of course, not many of the rodent-dispersed Conebushes are common or widespread, usually occurring in small little pockets in odd places. May this habitat rarity be the cost of being dispersed by seed predators?

Have a look at Who is pollinating Pr humiflora.

Back PAN 54