Premise Many traits covary with environmental gradients to form phenotypic clines. While local adaptation to the environment can generate phenotypic clines, other nonadaptive processes may also. If local adaptation causes phenotypic clines, then the direction of genotypic selection on traits should shift from one end of the cline to the other. Traditionally genotypic selection on non-Gaussian traits like germination rate have been hampered because it is challenging to measure their genetic variance.

Methods Here we used quantitative genetics and reciprocal transplants to test whether a previously discovered cline in germination rate showed additional signatures of adaptation in the scarlet monkeyflower (Mimulus cardinalis). We measured genotypic and population level covariation between germination rate and early survival, a component of fitness. We developed a novel discrete log-normal model to estimate genetic variance in germination rate.

Results Contrary to our adaptive hypothesis, we found no evidence that genetic variation in germination rate contributed to variation in early survival. Across populations, southern populations in both gardens germinated earlier and survived more. 

Conclusions Southern populations have higher early survival but this is not caused by faster germination. This pattern is consistent with nonadaptive forces driving the phenotypic cline in germination rate, but future work will need to assess whether there is selection at other life stages. This statistical framework should help expand quantitative genetic analyses for other waiting-time traits.

We measured the time to germination of Mimulus cardinalis seeds grown in the greenhouse and its effect on winter survival in field experimental gardens. See the main text for further explanation. We removed individuals that did not germinate or were suspected of contamination. Data on survival are missing from individuals that died because of transplant shock before the first winter.

The data are in a tab-delimited format that should work with most platforms and software.