Plant defences against mammalian herbivores : are juvenile Acacia more heavily defended than mature trees ?

Juvenile trees are expected to be more heavily defended against browsing mammals than mature plants. Juvenile and mature trees o f Acacia tortilis and A. nilotica occurring at Nylsvley, Northern Transvaal, were quantitatively compared in terms o f some potential chemical and physical defences. Neither species showed any significant difference between juvenile and mature trees in terms o f total polyphenol content, condensed tannin content, protein precipitating ability or protein content in leaves. Both species showed age-class differences in spinescence. In A. nilotica, thorns on branch tips were longer and more closely spaced and leaves were smaller in juveniles than in adults. Hence juveniles o f this species appear to be physically more heavily defended than mature plants. In A. tortilis, curved thorns were longer, but straight thorns were shorter than in mature trees. There was no difference between age classes in overall thorn density, but juveniles had a higher curved to straight thorn ratio. It is not obvious what the effects o f these differences might be on mammalian browsers.


INTRODUCTION
Woody plants that are browsed by mammals when still juvenile may show retarded growth and increased time to reach reproductive maturity (Bryant el al. 1983: Bryant el al. 1991b).Hence, woody plant defences against mam malian herbivores could be expected to be expressed more strongly in juveniles than in mature trees (Bryant & Kuropat 1980;Bryant el al. 1983;Bryant el al. 1991a. b. c).This has been documented in Alaskan paper birch (Betula resinifera) browsed by snowshoe hare (Lepus americanus) (Reichardt el al. 1984) and quaking aspen (Populus tremuloides) browsed by beaver (Castor cana densis) (Basey el al. 1990).The objective of this study was to assess whether some potentially defensive traits were expressed more strongly in juvenile Acacia trees than in mature trees.No studies comparing anti-herbivore defences of juvenile and mature Acacia trees have yet been reported.
There is circumstantial evidence for longer thorns being more effective as defences against browsers than shorter thorns in several Acacia species (Foster & Dagg 1972;Young 1987;Milewski et al. 1991).Cooper & Owen-Smith (1986) noted that thorns were more effec tive in restricting leaf losses to browsers when leaves were small.In Acacia karroo, the rate of intake of browse by goats was positively related to leaf mass per unit length of shoot (Teague 1989a).Thus leaf size and thorniness might act together in defence of trees against browsing ruminants.A. nilotica has paired, straight thorns, while A. tortilis has both 'hooked' thorns and straight thorns.

MATERIALS A N D M ETHODS
The study area was the Nylsvley Nature Reserve in northern Transvaal, South Africa (24° 39' S, 28° 42' E).Acacia tortilis and A. nilotica occur in the woodlands flanking the floodplain and in disturbed sites of former human habitation (Coetzee et al. 1976;Cooper & Owen-Smith 1986).
Ten pairs of trees comprising a juvenile and a mature plant no more than 50 m apart, were sampled for each species.Sampling was restricted to the disturbed Acacia woodland, and samples were paired to control for spatial differences in soil nutrients and browsing pressure.Adult trees were defined as those showing evidence of reproduc tive maturity (i.e.presence of flowers or seed pods).No trees taller than 1.7 m were considered for the juvenile class.
All collections took place on two consecutive days in March 1992, because during late summer all leaves are mature but not yet senescent.For each plant, leaves were taken from the terminal 250 mm of five branches on the northern side of the canopy at a height of between 0.75 and 1.7 m (within the browsing reach of impala and/or kudu).All leaf collecting took place between 06h00 and 09h00.Leaves of each pair of juvenile and mature trees were collected consecutively, with no more than 15 minutes between collection for members of any pair, to eliminate the effects of short time-scale fluctuations in leaf chemistry.Analysis of leaf nitrogen content provides a check for differences in leaf age between the juvenile and mature trees sampled.
Collected leaves were placed in plastic bags, and kept in a cooler box with ice blocks until the end of the col lecting session (maximum 3 hours).At the end of the ses sion, leaves were frozen in a deep freeze at -4° C.They were kept frozen until used for weighing and extraction for chemical analysis.

Physical defences
The length of all thorns within 250 mm of the branch tip on the same branches from which leaves were col lected was measured to the nearest millimetre.The density of thorns was measured by counting the number of thorns or thorn pairs within 250 mm of the branch tip.
The frozen leaves from each tree were weighed to the nearest 10 g.From these measurements, the mean fresh leaf mass for each tree was calculated to give an index of leaf size for each tree.The leaf matter used in extraction and for protein analysis (about 2.0 g) was weighed out from frozen leaf material.Several leaves were later dried in order to determine the relationship between dry mass and the mass of fresh leaf material so that chemical parameters could be reported in terms of 'percentage of dry mass'.

Chemical analyses
Phenolics were extracted according to the method described by Hagerman (1977).Approximately 2.0 g leaf matter was frozen with liquid nitrogen and ground in a mortar.The crushed material was then centrifuged three times at 2500 rpm using 20 ml of 70% acetone as solvent in each iteration.The supernatant was poured off each time and stored at 4° C until needed for chemical analysis.
Condensed tannin content was measured using the acid butanol assay (Porter et al. 1986).The absorbances at 550 nm were standardised against purified sorghum tannin.The total polyphenol content was determined using the Prussian blue assay (Price & Butler 1977).Absorbance was read at 720 nm.The standard was commercial tannic acid made up from powder form in the laboratory with 70% acetone.
The protein precipitating ability of the chemicals in the leaves was determined by means of radial diffusion using the method described by Hagerman (1977).The diameter of the ring of precipitated BSA in an agarose gel was measured, the square of the diameter being proportional to the amount of tannin in the sample (Hagerman 1977).
Protein content was estimated by determining nitrogen content by Kjeldahl oxidation using sulphuric acid and a selenium catalyst (Keeney 1982).The amount of am monium in the digest was determined colorimetrically fol lowing the method of Cataldo et al. (1975) and absorbance was read at 655 nm.The absorbances were compared to a standard curve plotted from the absorbances of dilutions of an ammonium sulphate solution.Crude protein content equals 6.25 x nitrogen concentration.
For each species, each variable was compared by means of a Wilcoxon paired summed-ranks test (Siegel & Castellan 1989) because it was not possible to confi dently assume normality in any of the parameters.The alpha level of acceptance for significant results was 5 per cent.Tests were one-tailed tests of the hypothesis that juveniles are more heavily defended than mature trees.

Chemical measures
The total polyphenol content of leaves of juvenile A. nilotica did not differ significantly from that of leaves of the mature plants (Table 1).The protein precipitating ability (as measured by radial diffusion) of the extracts from juvenile and mature A. nilotica did not differ sig nificantly, and neither did the protein contents of juvenile and mature plants.A. nilotica did not show any sign of containing condensed tannins, in that the extracts did not tum purple (indicative of condensed tannin) in the acid butanol test.
The polyphenol and condensed tannin contents for juvenile A. tortilis trees did not differ significantly from those of mature plants, nor did crude protein content or protein precipitating capacity (Table 1).

Physical parameters
Thoms of juvenile A. nilotica were significantly longer (F = 0.042) than those of mature trees (Table 2).The thorns of juveniles were, on average one and a half times as densely spaced as those of mature trees due to shorter intemode lengths (F = 0.003).Leaves of juvenile A. nilotica weighed only 75 % of the mass of those of mature trees (P = 0.003).
Comparison of the physical defences of juvenile and mature A. tortilis trees was complicated by the fact that this species bears both short, recurved ( "hooked' or 'curved') thorns and long, 'straight' thorns.The overall density of thorns on juvenile trees did not differ sig nificantly from that on mature trees (Table 2).
The ratio of straight to curved thorns was significantly higher in juveniles (P = 0.016) than in mature plants.Curved thorns were significantly longer on juvenile than on mature trees (P = 0.003), whereas straight thorns were longer on mature trees than on juveniles (P = 0.05).Leaf mass did not differ significantly between juveniles and mature trees of this species.

Chemical defence
The results of this study indicate that for both Acacia nilotica and A. tortilis, juvenile plants did not differ from those of adult trees in total polyphenol, condensed tannin or crude protein content.Hence juveniles of these two Acacia species did not appear to be more heavily defended chemically, at least at the time of year when samples were collected.This is in contrast to the findings of Reichardt et al. (1984) with regard to surface resins on twigs of paper birch.This may be related to differences in the way leaves and twigs are defended.Bryant et al. (1992: 345) note that i n every case that has been studied, the low palatability of the juvenile phase has been related to in creased concentrations of antifeedants in intemodes' (our emphasis).
Carbon-based secondary metabolites, such as tannins, would be costly defences during periods of rapid growth such as the juvenile stage (Bryant et al. 1991b(Bryant et al. . c. 1992)).Juveniles of both of these Acacia species show high in trinsic growth rates (Bryant et al. 1989).
It is also possible that polyphenols, condensed tannins and protein precipitating compounds do not play a defen sive role in these species, since Owen-Smith & Cooper (1987) classed both species as palatable to browsing ruminants.Our results confirm that A. tortilis has relative ly low concentrations of condensed tannins in leaves, and that A. nilotica has high total polyphenol contents, but no measureable condensed tannins, as found by Cooper & Owen-Smith (1985) and Cooper et al. (1988).Further more.our findings indicate that despite having a high polyphenol content.A. nilotica leaf extracts have a protein precipitating ability that is scarcely higher than that of A. tortilis.This suggests that the polyphenols found in the leaves of A. nilotica are mostly not tannins or other protein binding compounds.

Physical defence
Juvenile A. nilotica specimens have significantly longer and more densely spaced thorns and their leaves Thoms retard mammalian browsing by restricting bite size (Cooper & Owen-Smith 1986;Teague 1989a), but the relative importance of thorn length and thorn density in defence is not clear.Heavy browsing can induce longer thorns in regrowth than are seen in unbrowsed plants (Foster & Dagg 1972;Young 1987;Milewski et al. 1991), suggesting that longer thorns might be more effective browsing deterrents.The measurements presented here in dicate that juvenile Acacia nilotica are physically better defended than mature trees only if two conditions hold.Firstly, it must be shown that greater thorn length and/or density provides better defence against browsers than shorter or less dense thorns.Secondly, the pattern shown by juveniles under one metre in height ( Cooper & Owen-Smith 1986) must be explained.The smaller leaves of juveniles in our study may assist in further restricting feeding rates of browsers.
In A. tortilis, curved thorns are less dense but longer, and straight thorns more dense but shorter, in juveniles than in mature trees.Leaf size did not differ between clas ses.Curved thorns retard biting rates of kudu and impala (Cooper & Owen-Smith 1986).The longer, less densely spaced curved thorns in the juvenile trees do not neces sarily indicate heavier juvenile defence because it is not known how (if at all) the length and density of curved thorns contribute to their effectiveness as defences.Dif ferences in the lengths and relative densities of both thorn types in Acacia tortilis cannot be said to support an hypothesis of heavier juvenile defence until the defensive roles played by the two thorn types is properly understood.