Interesting study! Thank you for pointing it! All urban sites were botanical gardens and parks (specially selected flowering plant rich places with high biodiversity), but rural sites were selected using GIS and land-use maps to be dominated by agricultural land and semi-natural/forest cover (hence having less biodiversity - the common agricultural practice is growing monocultures). I think that could have larger impact on the size of bumble bees than different fragmentation level of urban and rural landscapes.
