You are viewing an archived copy of this website captured Mon Nov 05 11:37:32 AEDT 2012

All Content © CRC for Forestry 2007

How far does Eucalyptus globulus pollen travel?

Martyn Lavery sampling

Martin Lavery from Arianda Pty. Ltd. assisted with the gene flow study by providing his climbing skills to sample seed from the upper crown in the continuous population of Eucalyptus globulus in the Strzelecki Ranges, Victoria.

Brad Potts
School of Plant Science
University of Tasmania

An understanding of the extent to which pollen travels in the landscape is important in assessing and managing the risk of gene flow between plantation and native forest. While we have some information on the pollen dispersal curve for the insect pollinated species Euclayptus nitens (Barbour et al. 2005), such curves have not been published for E. globulus which has much larger flowers and is pollinated by birds as well as  insects (Hingston and Potts 1998; Hingston et al. 2004).

In a collaborative study lead by Dr. Makiko Mimura from the University of Tsukuba (Japan) and started in 2006, molecular markers were used to model effective pollen dispersal in native populations of Eucalyptus globulus.  Numerous factors may affect pollen dispersal, but landscape characteristics such as tree density are believed to be very important.  We studied two pairs of native populations – a highly fragmented population in an agricultural landscape and a relative dense, continuous population in both Tasmania (Dennes Point and Apollo Bay on Bruny Island) and Victoria (Toora and Strzelecki Ranges).  The Victorian populations are located inland and are separated by approximately 35 km while the Tasmanian populations are located near the coast and are only 9 km apart.  Using eight microsatellite markers, we genotyped 1289 open pollinated offspring collected from 20 trees in each population.  Using these data we were able to estimate contemporary mating patterns, including outcrossing rates and the pattern of effective pollen dispersal.  Within each population, the parent trees were separated by a range of distances, and pollen dispersal curves were estimated by studying the decay in correlated paternity between pairs of parent trees as the distance between the pairs of trees increased, using the software KINDIST.

There are few pollen dispersal studies in the literature which have such landscape level replication, but such replication is important to provide robust estimates of the pollen dispersal curve and understand its variability.  The mating patterns in the two continuous populations were similar, despite large differences in the density of trees.  In contrast, the two fragmented populations were variable and idiosyncratic in their mating patterns, particularly in their pollen dispersal curves.  The continuous populations showed relatively high outcrossing rates (86-89%) compared to the fragmented populations (65-79%).  A greater proportion of trees contributed to reproduction in the fragmented compared to the continuous populations.  Pollen dispersal was best modeled with an exponential power function in all populations, which in three of the four populations indicated that the density of dispersed pollen was relatively low beyond 50 m (3-4 times the canopy height) from the parent tree.  While long distance dispersal is difficult to quantify, these functions suggest a long ‘fat-tail’ of rare longer distance dispersal events.  The exception to this pattern was the fragmented population in Tasmania at Dennes Point where the functions suggested a relatively greater proportion of medium distance pollen dispersal events as far as 100 m from a parent tree.  In fact, there was a trend for effective pollen dispersal to be increased in the fragmented compared with the continuous landscape, due to greater medium distance dispersal at Dennes Point but slightly more frequent long-distance dispersal in the more fragmented landscape at Toora.

Mimura M, Barbour RC, Potts BM, Vaillancourt RE and Watanabe KN (2009). Inbreeding and pollination distance in continuous and fragmented forest landscapes. Molecular Ecology 18: 4180-4129. [View article]

Biobuzz issue nine, august 2009