The work in Glacier Bay, Alaska, re-discovering and measuring the 1916 William Cooper plots, has been accepted for publication in Ecology. This is an important dataset, 100 years of fairly regular measurements of vegetation composition and characteristics. It is truly unique, not only due to its age, but because of its ability to shed light on primary succession and chronosequences writ large, which make a large assumption that space can be substituted for time.
The 100 year record in Glacier Bay suggests that those assumptions are not necessarily valid, even in a near ideal condition - low species diversity, relatively strong abiotic filters (read: harsh weather), and consistent year to year climate. Contingency, primarily driven by the founder/first come first served effect, seems to dominate species composition. However, metrics of diversity, which ignore specific species in favor of overall community characteristics, were better behaved - species evenness, for example, decreased in a fairly linear pattern across time regardless of plot location.
There are more results to come out, including a more detailed soil assessment and inclusion of new data from the 1960s to 1980's to fill in a hole in the dataset currently, but these initial results will be available from Ecology soon.