Bernhardt & Weston, The pollination ecology of Persoonia 
781 
Nectar samples were collected by bagging individual flowers or whole inflorescences 
in situ overnight. The following morning each flower was probed with a ten lambda 
microcapillary tube. Since no flower produced as much as 10 microlitres of nectar, 
samples represent combined measurements from several flowers on the same shrub or 
several shrubs of the same species. Once ten microlitres was obtained the contents of 
the tube was deposited on Whatman's Filter Paper No.l. The contents were air dried, 
labelled and stored in a paper envelope and mailed to C.E. Freeman (Dept, of Biological 
Sciences, University of Texas, El Paso) to identify component sugars and record their 
relative proportions. 
Analyses of foraging insects 
Observation of prospective pollinators and analyses of the pollen they carried followed 
Bernhardt (1984, 1995). The behaviour of insects on Persoonia flowers was recorded 
from 9 am until 4:30 pm. Insects were collected only if they were observed probing for 
nectar or actively collecting pollen. Insects were killed in jars containing fumes of 
ethyl acetate. Insects caught on different Persoonia species were always killed in separate 
jars. Jars were cleaned after each collecting trip to avoid contamination upon reuse. 
To analyse pollen carried by insects, each insect was placed on a clean glass slide and 
'bathed' in a couple of drops of 100% ethanol. When the ethanol evaporated, the 
residue remaining on the slide was mounted in two or three drops of Calberla's fluid 
(Ogden et al. 1974). Identification of pollen was made under light microscopy. However, 
since more than one insect was killed in the same jar, some pollen contamination of 
insect bodies was possible. Therefore, a pollen taxon was not recorded as present on 
an insect unless more than 25 individual monads or 25 individual polyads (e.g. of 
Epacridaceae) could be counted under each cover slip (see Bernhardt 1984, 1995). 
Light microscopy showed that, as under SEM (Feuer 1986), pollen grains of different 
Persoonia species may be identified using a combination of characters including the 
physical size of the grain, the length and angle of pollen lobes, the inflation of pore 
opercula, density of tectum scabs and the frequency of tetraporate grains and/or 
irregular lobes. While these characters intergrade broadly between many species it is 
possible to discriminate between the pollen of up to three, co-blooming, sympatric 
Persoonia species carried on the same insect and washed onto the same slide. 
The length of each insect specimen was measured from its labrum to the apex of its 
abdomen. The insect was pinned, labelled to cross-reference with its pollen slide and 
sent to Dr K. Walker (National Museum of Victoria, Abbotsford) for identification. 
Results 
Floral phenology 
Herbarium records indicated that there are likely to be some populations within the 
genus Persoonia in bloom each month of the year within New South Wales. However, 
flowering is greatest from December through April (Table 2). Interspecific overlap of 
flowering periods was found for each Persoonia species. 
The flowering of a few species, such as P. laurina and P. chamaepitys was found to 
peak between late spring and early summer. Collections were few but the flowering 
of montane and subalpine P. subveliitina and P. arborea appeared confined to late 
summer. Flowering on a stem is acropetal to subacropetal in all species studied. 
