I am trying to compare the dispersal effectiveness of Chinese yew between different geographical regions.
I have collected the seeds foraged by birds and their post-foraging perching behaviors at each region.
Does anybody have any good suggestions for comparing dispersal effectiveness?
Dispersal effectiveness was estimated considering aspects of the quantity and quality components of seed dispersal:
-Quantity: Estimated using data on the relative abundance of birds and the frequency of visits to the fruits per hour.
-Quality: Estimated using data on seed germination (after their removal by birds)
and the probability of seeds being deposited across different region.
If you need to estimate only the quality you must determine the proportion of germinated seeds in the deposit site (the criterion used to consider a seed
germinated was radicle emergence) and ther germination rate in each region.
You may use the ANOVA (one way) to find significant differences on the quality across different region, supplemented by the post-hoc test, Tukey.
Any criticism or comments are welcome.
Hello Ning,
i recommend that you read the paper on the following link:
http://onlinelibrary.wiley.com/doi/10.1111/j.1469-8137.2010.03402.x/full
Good luck
Dear Ning Li
I would say - the stats method depends on how you measured / define the seed dispersal effectiveness. If you are interested in the dispersal of particular species of particular vegetation (set of species) - an empirical method is eg setting up a seed traps (eg astroturf - standardized size patches of artificial grass) in various distance from source populations or in a sites that are of interest for you. I would additionally account for 1. number of seeds (purely stochastic effects) 2. size of seeds (smaller seeds generally travel further) 3. a seed characteristic - adaptations to float on water / fly with wind etc. (from databases on seed traits or from seed atlases).
greetings
Agata
Pactli makes an excellent point, that quantity and quality are two different properties that should be assessed differently.
I would not use ANOVA for assessing quality (rate of germination) because we are dealing with an underlying binary variable (a seed did or did not germinate) which implies a binomial distribution, making a generalized linear model with a binomial distribution and logit link function more appropriate. Some logistic regression procedures can handle anova-type variables (i.e., categorical variables, also called class variables, or dummy variables - region would be such a variable). Under some situations the binomial approaches the normal distribution, but you cannot assume that the normal is a good substitute for the binomial without checking.
I would add a possible third variable to quantity and quality, and that is the spatial distribution of dispersed seeds. These distributions are often exponential (decreasing exponential) out from the mother plant. With bird dispersal, however, more likely the seeds are dispersed in clusters around perching trees, perhaps with an exponential decline in density between perching trees as you get further away from the mother plant.
Norma Fowler
I agree, first define SDE because there are so many different definitions!
Andrew :-)
Here is the best definition of SDE I could find: Seed dispersal effectiveness (SDE) is the contribution of dispersers to plant recruitment and is estimated as the product of the number of seeds dispersed (quantity) and the probability of recruitment of each dispersed seed (quality). And this paper compares SDE: DOI: 10.1890/14-0655.1.
I hope this helps!
Andrew :-)
Check my new paper
Article Seed dispersal effectiveness by frugivorous birds. Identifyi...
All the best
- Badges
- Science topic
- Similar topics
- Biological Science
- Biota
- Eukaryota
- Fungi