Is landscape connectivity a dependent or independent variable?

With growing interest in landscape connectivity, it is timely to ask what research has been done and what remains to be done. I surveyed papers investigating landscape connectivity from 1985 to 2000. From these papers, I determined if connectivity had been treated as an independent or dependent variable, what connectivity metrics were used, and if the study took an empirical or modeling approach to studying connectivity. Most studies treated connectivity as an independent variable, despite how little we know about how landscape structure and organism movement behaviour interact to determine landscape connectivity. Structural measures of connectivity were more common than functional measures, particularly if connectivity was treated as an independent variable. Though there was a good balance between modeling and empirical approaches overall - studies dealing with connectivity as a dependent, functional variable were mainly modeling studies. Based on the research achieved thus far, future landscape connectivity research should focus on: ( 1 ) elucidating the relationship between landscape structure, organism movement behaviour, and landscape connectivity ( e. g., treating connectivity as a dependent variable ), ( 2 ) determining the relationships between different measures of connectivity, particularly structural and functional measures, and ( 3 ) empirically testing model predictions regarding landscape connectivity.