Australian Weeds Vol. 2(4) Winter 1983 127

generally successful; less than 2"70 of shoots sampled eitheT carTied a ny Howers or showed ev id ence that ripe RESEARCH REPORTS fruit had fallen. Even in this minority of shoots, losses did not exceed 20"70, so eTTOTS due to fruit loss or incomplete fruit se t were negligible compaTed with the very hi gh natural variation among shoots. The numbeT of immatuTe fTuit formed was treated as fTuit production. FaT each shoot the following data were collected: diameteT of the stem at 20 em from the ground; two diameteTs of the canopy outline, at right-angles Fruit production by two species of privet, to one a notheT, from which the aTea of Ligustrum sinense Lour. and L. lucidum W.T.Ait., the canopy was calculated as an elli pse in Sydney wi th the two measured diameters; the height of the canopy at its highest point; an estimate of the percentage M. Westoby, J. Dalby and L. Adams-Acton projective foliage coveT (Specht, 1970) School of Biological Sciences, Macquarie Un iversity. North Ryde, due to other plants aT shoots, of any New South Wales 2113 species, over the shoot in question; an estimate of the numbers of fruit carried on the stem. This could not be made by complete census since the numbeTs Summary which Buchanan (1978) showed was fTequently exceeded 10' per shoot. readily eaten by biTds and dispersed up Fruit production by Chinese privet The estimate was made in the field by I (Ligustrum sinense) and large· leaved to km. This report quantities the fruit counting fruit numbers on sample­ production of both pTivet species as a privet (L. lucidum) was studied at 58 lengths of about 20 cm of fruiting function of some aspects of plant size sites in urban bushland in the northern bTanch and extTapolating the numbeT and illumination of the canopy at a suburbs of Sydney in 1978, and again peT unit length to the whole bTanch. number of sites in the nOTthern sub­ at eight of the sites for L. sinense in At 35 sites samples of soil from the uTbs of Sydney. 1979. L. sinense averaged 1300 fruits sUTface 5 em were coll ected; Ligustrum m" of canopy and the less widely dis· spp. have many roots in the sUTface tributed L. lucidum averaged 400 m·'. Methods soil. A feTtility index was obtained for Fruit production in both species was .... Both Ligustrum species fTequently these samples by bioassay. A tomato duced with increasing shade. Fruit pro­ sucker from the base, and multi­ seedlinr was grown in each of two Tep­ duction per unit area of canopy was shooted individuals are often mOTe licates of each soil in 10 em pots in the glasshouse. AfteT 18 days fTom emer­ less for smaller diameter than for larger common than those of tTee form. It is diameter shoots in L. sinense but was commonly not possible to tell which gence tops and roots were harvested not related to shoot diameter in L. Iuci­ shoots are part of a given clone, so we and dTied, a nd the Tela tive growth rate (g g" day") from emeTgence to haTvest dum. In both species there was great used the shoot ratheT than the genetic variation in fruit production between individual as the unit of study. The was used as an index of fertilit y. shoots of similar sizes, Bnd also con­ shoot was defined as a plant aT part of siderable variation in fruit production a plant which had a single stem at Results between sites, which was not correlated 20 em fTom the ground. with soil fertility as measured by a A total of 58 sites weTe visited once Differences between species crude bioassay. Sites with high L. sin­ each between 29 ApTil a nd 10 July, Ligustrum sinense was more common ense fruit production in 1979 were not 1978 at various locations in the north­ wi thin the aTea than L. lucidum. At 20 the same as those with high production ern suburbs of Sydney. At each site 10 sites L. sinense was the only species in 19~8. to 30 shoots of each privet species present, while L. lucidum occurred present were examined. In addition, 141 alone at only one site; both species were shoots were examined along road pTesent at 37 sites, generall y with Introduction verges or in private gardens, making a L. sinense being much mOTe a bundant The privets Ligustrum sinense Lour. total of 363 shoots of L. lucidum and than L. lu cidum. Both species weTe and Ligustrum lucidum W.T. Ail. are 1127 of L. sinense. At each site shoots found on remaTkably vaTied soil s, among the more important urban were chosen to Tepresent a Ta nge of Tanging from neaTly pure sands weeds of the Sydney area (Bradley and sizes, but were not selected in any other through friable loa ms to nearly pUTe Bradley, 1967). They are abundant both way. In May and June 1979 eight of the clays, a nd there was no diffeTence be­ in gardens and in urban bushland, sites studied during 1978 (aTbitrarily tween the species in the fertilit y index forming very dense stands a nd appaT­ chosen) weTe Tevisited and 20 shoots of of sites wheTe they occuTred (L. sillellse ently excluding native species. Both L. sinense examined at each to com­ 0.187 ± 0.058 vs. L. lucidum 0. 185 ± local authorities and amateuT bush­ pare fruit production in consecutive 0.061; t = 0.11, df = 60, P>0.5). protection organizations now make yeaTS. The average properties of the effoTts to contTol them. Adamson In both years we aimed to sample sampled shoots of both species over a ll (1977) suggested that a factoT con­ shoots while they caTried immatuTe si tes a re summaTized in Table I. It tributing to theiT success and Tapid fruit, afteT fruit set but befoTe fTuit should be noted that these stems we Te spread is a heavy production of fruit, ripened and fell aT was eaten. This was not a fully randomized sample wi th re- 128 Australian Weeds Vd . 2(4) Winter 1983

Tablel Average properties of sampled shoots of Liguslrum lucidum and .05). The data suggest that for L. sin­ L. sinense in bushland in northern Sydney ense the response takes the form of de­ pressed fruit yield in the smallest stems Species (Figure I). Characteristic L. lucidum L. sinense Fruit production per unit canopy Meon S.E. Mean S.E. area was reduced gradually and pro­ gressivelyas percentage cover over the total fruit numbers per shoot 2300 360 2800 340 canopy (broadly speaking, the degree fruit numbers per square metre of canopy· 400 45 1300 130 stem diameter (cm at 20 em)· 9.9 0.30 3.4 0,07 of shading) increased (Figure 2); full 1 canopy area (m ). 5.7 0.25 2.4 0.09 cover reduced fruit production by canopy height (m)* 4.6 0.07 2.7 0.04 50"7. to 70% compared with full light percentage cover over canopy· 33 1.5 42 0.95 in both species. Since smaller shoots were generally Significant difference between species is indicated by • (P<.OOI). more shaded, the apparent effect of shade might actually have been an effect of shoot size or vice versa. An spect to size. However, far more small dum, with the result that average pro­ analysis of covariance of these factors stems were left unsampled for L. sin­ duction per stem was similar in the two (Table 2) showed that shoot size and ense than for L. lucidum. The tendency species. shading explained substantially differ­ of L. lucidum to have larger stem dia­ ent parts of the variance in fruit pro­ meter, canopy area and height is there­ duction per square metre for L. sin­ fore certainly real, while the tendency Factors affecting fruit production per unit canopy area ense, but for L. lucidum the weak cor­ of L. sinense to be slightly more shaded relation of fruit production per unit than L. lucidum may have been a pro­ Fruit production per unit canopy area canopy area with stem diameter ap­ duct of its generally smaller size. Ligus­ was positively correlated with stem dia­ pears to have been a side-effect of the meter in L. sinense (r 12 .08, P < Irum sinense produced more fruit per = lesser shading of larger shoots . unit area of canopy than did L. luci- .01), and in L. lucidum (r12 = .133, P< Differences between sites and years There was substantial variation be­ tween sites in mean fruit production per unit canopy area, both in 1978 and in 1979. This variation was not cor­ related with the fertility index of the sites (r2 = .09, n = 27 for L. luci­ dum, r' = .01, n = 34 for L. sinense), nor with any other feature of the sites 12 6 that we observed. The sites with heavy 2400 fruit production were not the same for the two species. When L.' sinense was resampled at 8 sites in 1979, there was no tendency for the same sites to have 2000 25 the highest fruit production per square metre of canopy in both years (F, = 1.75, df = 7, 251, P > O.lO). For these reasons we believe the variation be­ 1600 tween sites was largely due to the varia­ tion between individual stems. This 39 variation was- very largt; as is common 44 0 of fruit production (Harper, 1977); for 1200 both species about 5 % of the shoots t sampled accounted for about 60% of the total fruit production observed.

800 496 t DIscussion In both species larger diameter stems produced more fruit per unit canopy 400 area. Nevertheless stems of less than 25 mm diameter at 20 em from the ground could produce substantial amounts of fruit. As a result, even if 0 ~~--~--L-~------~2~0 ~0 ------all stems in an area were removed, it 25 50 75 100 would seem that new seedlings derived Slem diameter (mm al 20 em) from the soil seed reserve could not be allowed to grow very large before re­ Flgu,. 1 Response oI lruil numbers produced per square metre 01 Ugustrum sinense (e) and L. lucidum (0 ) plenishment of the soil seed reserve 10 stem diameter (mm at 20 em Irom the ground). Bars indicate t one standard error; numbers show sample size IOf each mean. would begin. Aus tralian Weeds Vol. 2(4) Winter 1983 129

2200 Overhead shading accounted for a reduction by more than 500/0 in fruit production. Leaving other tree and 2000 shrub species in place durin& a control programme would contribute to re­ ducing fruit production, and it might also be useful to leave large killed 1800 plants of Ligustrum in place. 198 It should be pointed out, however, that the importance of high fruit pro­ 1600 duction to the growth and spread of 361 privet populations is not known. Clearly some fruit production is essen­ 1400 tial to spread; but to what extent reducing fruit production by less than 100% would slow spread or recoloni­ zation of cleared areas is unknown. 147

Acknowledgements This work was supported by the New South Wales State Pollution Control 169 Commission.

252 References Adamson, D. (1977). Privet and the prob­ 600 ~1l 9 lem of control of exotic plants in urban bushland. Submission to Noxious Plants Advisory Commission of New South Wales, II pp. 400 ~ 121 Bradley, E. and Bradley, J. (1967). Weeds 27 and their control. Mosman Park lands and Ashton Park Association. Sydney. 200 Buchanan, R. (1978). Pied currawongs (Slrepera graculina Shaw): their diet and weed dispersal in some urban bushland, Sydney, N.S.w. Bulletin of the Ecological Society of Australia 8(1):6. 0 ~-----7~----~~----~~----~------~ o 20 40 60 80 100 Harper, J. L. (1977). Population Biology of Percentage cover overhead Plants. Academic Press, New York. Specht, R. L. (1970). Vegetation. In The Figure 2 Response of fruit numbers produced per square metre of Ligustrum sinense (- ) and L /ucidum (0) Australian Environment ed. G. W. to shading, measured as percentage projective foliage cover of all species over the shoot. Bars indicate ± one standard error; numbers show sample size for each mean. Overall correlation for L. sinense Leeper. Fourth edition. Melbourne Uni­ '1 2 - -.09, P

Table 2 Analysis of covariance of the effect of stem diameter and percentage cover overhead on fruit production per square metre of canopy for two Ligustrum species MANUSCRIPTS REQUIRED Source Percent sum of squares lnkata Press would like to hear from L. sinense L. lucidum authors in agriculture and the bio­ logical and earth sciences. The saturated model material submitted need not be in an (stem + shade + interaction) 1.5··· 4.3" advanced state of preparation as all additive model ideas or suggestions for new books (stem + shade) 1.4·· 4.1" stem, adjusted for shade 0.60·· 0.09 will be received with interest. shade, adjusted for stem 0.59" 3.7··· lnkata Press Ply Ltd , interaction 0.12 0.14 4 LDngbourne Avenue. residua] 98.5 95.7 Clayton North. Victoria 3168. Corresponding F-ratios are significant when indicated by"· (P<.OOI) or •• (P<. OI).