Aim 

　　Effort in collecting biodiversity information often varies strongly in space and may be driven by environmental, cultural and socio-economic factors. Understanding the constraints on collecting effort is crucial for identifying potential bias in distributional databases and for making future surveys more efficient. Here we test six competing hypotheses on drivers of geographical variation in collecting effort and identify the main factors shaping the geography of floristic collections in China. 

Location 

　　China. 

Methods 

　　We used the most comprehensive database of Chinese vascular plant distributions including 4,338,516 county-level occurrences derived from herbarium specimens and literature sources. Explanatory variables were assembled representing six different hypotheses: accessibility, human population density, the ‘botanist effect’, mountains, water availability and conservation priority. Ordinary least-squares models with eigenvector-based spatial filters were applied to investigate their effects on spatial patterns of two different facets of collecting effort, i.e. collection density and inventory incompleteness. 

Results 

　　All hypotheses except accessibility and human population density received significant support. Elevational range was the strongest predictor with a positive effect on collection density. Inventory incompleteness in turn was best predicted by human population density, but unexpectedly showed a positive effect. In contrast to previous studies, collecting effort was only weakly and negatively related to road density. Counties with herbaria had significantly higher collecting effort, and the presence of herbaria had weakly positive effects on neighbouring counties. 

Main conclusions 

　　Our results indicate that China's mountains are most intensively and completely collected, whereas densely populated areas are surprisingly under-sampled. Because densely populated areas are more seriously threatened by land-use change, our results show a need to increase biological collections in those areas for conservation assessment and monitoring. More generally, our study suggests that collecting effort and its environmental and socio-economic constraints have a strong region-specific component influenced by cultural context and by different botanical traditions.