Premium
Effects of within‐plant variability in seed weight and tannin content on foraging behaviour of seed consumers
Author(s) -
Shimada Takuya,
Takahashi Akiko,
Shibata Mitsue,
Yagihashi Tsutomu
Publication year - 2015
Publication title -
functional ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.272
H-Index - 154
eISSN - 1365-2435
pISSN - 0269-8463
DOI - 10.1111/1365-2435.12464
Subject(s) - biology , tannin , foraging , trait , wood mouse , botany , selection (genetic algorithm) , condensed tannin , horticulture , ecology , proanthocyanidin , apodemus , polyphenol , biochemistry , antioxidant , artificial intelligence , computer science , programming language
Summary Individual plants provide a habitat patch for foragers where the sought‐after resources (e.g. leaves, fruits and seeds) are clustered in locally dense aggregations. Characteristics of these resources often vary greatly even within individual plants, which is known as within‐plant or subindividual variation. To best describe properties of the patch (individual plants), the higher moments of trait‐value distributions must be included in addition to the mean values. However, the question whether differences in within‐plant variability of a given trait influence the foraging behaviours of consumers has been mostly untested. To test the hypothesis that differences in within‐plant variation of traits can influence the patch selection behaviour of seed consumers, we used the mean, coefficient of variation (CV) and skewness of within‐plant distributions of two Quercus serrata (Thunb) seed traits (weight and tannin content) as descriptors of plant phenotypes. We then examined the relationships between these descriptors and patch selection by the wood mouse Apodemus speciosus , where individual Quercus trees formed a foraging patch for seed removal. We collected 8594 seeds from 26 individual trees in a forest in northern Japan. We then measured their weight and tannin content, marked and returned them to the locations where they were found, and traced their fates in relation to seed removal by the wood mouse. Tannin content was nondestructively estimated using near‐infrared spectroscopy (NIRS). There was notable within‐plant variation in seed weight and tannin content. Generalized additive modelling revealed that trees with a large mean and CV of seed weight and those with a small CV and skewness of tannin content had a high frequency of seed removal. These results highlight the importance of considering the within‐plant distribution of seed trait values, in addition to the mean trait values of individual seeds, when describing plant phenotypes. The inclusion of these data is essential to understand the ecological and evolutionary processes at work in plant–animal interactions.