Reduced precipitation lessens the scaling of growth to plant N in mesic grasslands

Grassland production is sensitive to both precipitation and plant N
accumulation and utilization, such that change in one variable influences
grassland response to the second variable.  We investigated
effects of interannual variation in precipiation on the response of
'community'-scale values of relative growth rate (RGR) to two
multiplicative components of RGR, nitrogen productivity (NP; rate of
change in biomass/plant N), an index of N utilization efficiency, and
plant N concentration ([N]), in two grassslands in Texas, USA. 
Grasslands included a planted mixture of perennial grass and forb species
and a monoculture of the perennial C4 grass Panicum vigatum that was
invaded by multiple plant species.  RGR and its N components were
measured at the spatial scale of 7-m diameter circular patches near the
spring peak in mixture biomss during each of 5 years.  We found
that RGR varied substantially among patches and years and between the
planted mixture and monoculture.  RGR variation was strongly
correlated with variation in NP.  Precipitation during the 3
months prior to RGR measurement mediated that RGR response to NP by
altering the correlation between NP and [N] in both grasslands. 
Reduced precipitation led to more negative NP-[N] correlation
coefficients, which reduced proportional change in RGR per change in NP by
as much as 30% even in the absence of a precipitation effect on means of
RGR and NP.  Our results highlight an under-appreciated aspect of
the pervasive role of precipitation in grassland growth that was mediated
via change in the growth benefit derived from plant N.