Additive genetic variance for lifetime fitness and the capacity for adaptation in an annual plant

The immediate capacity for adaptation under current environmental conditions is directly proportional to the additive genetic variance for fitness, VA(W). Mean absolute fitness, , is predicted to change at the rate , according to Fisher’s Fundamental Theorem of Natural Selection. Despite ample research evaluating degree of local adaptation, direct assessment of VA(W) and the capacity for ongoing adaptation is exceedingly rare. We estimated VA(W) and in three pedigreed populations of annual Chamaecrista fasciculata, over three years in the wild. Contrasting with common expectations, we found significant VA(W) in all populations and years, predicting increased mean fitness in subsequent generations (0.83 to 6.12 seeds per individual). Further, we detected two cases predicting “evolutionary rescue”, where selection on standing VA(W) was expected to increase fitness of declining populations ( < 1.0) to levels consistent with population sustainability and growth. Within populations, interannual differences in genetic expression of fitness were striking. Significant genotype-by-year interactions reflected modest correlations between breeding values across years (all r < 0.490), indicating temporally variable selection at the genotypic level; that could contribute to maintaining VA(W). By directly estimating VA(W) and total lifetime , our study presents an experimental approach for studies of adaptive capacity in the wild.


43
A population's mean absolute fitness, W ̅ , in terms of the per capita contribution of 44 offspring, corresponds to its growth rate and is thus a measure of its degree of adaptation (Fisher 45 1930, pp. 25, 37;Roughgarden 1996, Ch. 4). Fisher (1930) showed that the immediate capacity 46 for further adaptation is proportional to the magnitude of a population's additive genetic variance 47 for absolute fitness, evaluated as an individuals' lifetime contribution of offspring to the 48 population (Fisher 1930;Price 1970Price , 1972Ewens 2004  in Iowa (41°40'44.2"N 92°51'24.9"W; hereafter "CERA"). 167 To obtain pedigreed populations of seeds from each of these populations, six seeds from 168 each maternal plant were surface sterilized using an 8.9% bleach solution followed by a 70% 169 ethanol rinse. We used 100 grit sandpaper to scarify the seeds, imbibed them in sterile water for 170 three days, and then planted them in small peat pots.

Fitness surveys 201
We regularly recorded fitness components for each planted seed throughout the growing 202 season of the following year, including emergence, survival to flowering, and reproductive 203 output (Table 1) Therefore, we retained these plants in censuses after herbivory to account for this late season 213 seed production. Because individuals ranged widely in number of fruits set (1-68), and fruits 214 explosively dehisce at maturity, we were unable to collect all mature fruits or count all seeds 215 each plant produced. Therefore, seed counts were obtained on a subsample of fruits produced. 216 Subsampling was accounted for in the statistical analyses (see below). scale using a mapping function that adds a component of a fixed effect (see Table 3   each population revealed large differences between years in family-specific expressions of total 294 lifetime fitness (Fig. 3). Breeding values exhibited small to moderate correlations between years, 295 with two cases of modest negative correlations: 2015 vs. 2016 in McCarthy Lake and 2016 vs. 296 2017 in CERA (Fig. 3). The findings of significant genotype-by-year interactions indicate that 297 families contributing disproportionately to the next generation often differ among years. 298

Mean fitness and predicted changes in mean fitness 299
The inclusion of block effects in all fixed-effects aster models for W ̅ explained significantly 300 more variation than models without (Grey Cloud and McCarthy Lake: all test df = 7, all test 301 deviance > 4.26, all P < 0.05; CERA: all test df = 3, all test deviance > 27.92, all P < 0.0001). 302 As with estimates of VA(W), estimates of mean fitness varied among populations and years 303 (Table 3;  fitness is consistent with population sustainability and growth of the progeny generations (Table  315 3; Fig. 4). 316

317
The immediate capacity for population-level adaptation is strongly dependent on the 318 presence and magnitude of additive genetic variance for lifetime fitness (Fisher 1930;Ewens 319 2004). Our study detected significant and substantial capacity for immediate adaptation through 320 current natural selection on standing levels of VA(W) in three populations in three years. Among 321 these nine cases, we detected two instances of populations declining numerically, and in those 322 cases obtained predictions of evolutionary rescue increasing mean fitness to levels consistent 323 with maintaining or increasing population size. Genotype-by-year interactions indicated that 324 genetic effects on fitness differed among years. These interactions reflect differences among 325 years both in the magnitude of VA(W) and in genotypic fitness rankings. Predicted change in 326 mean fitness also varied among years, suggesting that the capacity for adaptation is strongly 327 influenced by temporal environmental variation. Below, we discuss the importance of these 328 findings and consider how our results from direct study of the adaptive process in the wild 329 illuminate the potential for ongoing adaptation. We also relate our findings to the potential for 330 evolutionary rescue of declining populations. Finally, we describe how our experimental and 331 analytical approaches overcome obstacles commonly associated with evaluating adaptive 332 capacity in the wild, and we advocate for the broader implementation of these approaches. 333

Expression of additive genetic variance for lifetime fitness 334
Our detection of prevalent and non-negligible additive genetic variance for lifetime 335 fitness, characterized as the number of seeds produced per seed planted, indicates differential 336 genetic contributions to the progeny generation and, via the FTNS, clear capacity for ongoing 337 adaptation under natural selection in each year. We estimated significant VA(W) in the home site 338 of all three populations in all three years of study (Fig. 2) Our finding of prevalent interactions between genotype and year can be partitioned into 352 two aspects of environmental dependence of genetic expression (Falconer 1952). First, 353 correlations between years of family-specific breeding values for lifetime fitness were generally 354 low and, in two cases, slightly negative (Fig. 3). Thus, genotypic contributions of offspring in 355 one year are not predictive of contributions in another year. Interestingly, the two strongest 356 genetic correlations between years were not for consecutive years. These modest between-year 357 genetic correlations suggest that the response to selection may not be accompanied by directional 358 change in the frequency of the same alleles over multiple years. Second, for each population, 359 estimates of VA(W) differed strikingly among years (Table 3) Fig. 2; Fig. 3). Given the short 380 timeframe of our study (i.e. single generation replicated over three years), the impact of novel 381 mutations on lifetime fitness would be negligible. Additional phenomena may also contribute to 382 the conservation of genetic variation (e.g., marginal overdominance; Levene 1953; Gillespie 383 1984), but the differential and variable genetic contribution of families to the following progeny 384 generation across years (Fig. 3) is consistent with a role for temporal environmental variation in 385 impeding depletion of VA(W) by natural selection. These annual variations may contribute to the 386 maintenance of the appreciable levels of observed VA(W) in our study. 387 Our measures of immediate capacity for adaptation can alternatively be expressed as 388 evolvability (sensu Houle 1992, see Table 3) Table 3) were sufficient to increase mean fitness under contemporary natural 400 selection. In the case of the largest estimate of VA(W) (Grey Cloud in 2017), contemporary 401 selection was predicted to drastically increase fitness to several times that of the parental 402 generation (Table 3, Fig. 4). 403 Our study identified two cases where evolutionary rescue was predicted to increase mean 404 fitness of declining populations to levels consistent with sustainability and even population Our assessments of fitness encompass components of fitness expressed across the entire life span 420 of individuals, from each seed planted through to the seeds it produced; these fitness evaluations 421 are thus uncommonly complete. Nevertheless, some aspects of fitness are not included. For 422 example, this study did not account for fitness realized through siring of seeds (male fecundity). 423 In a companion study, however, Kulbaba and Shaw (in review) found positive genetic 424