Holder and Field: Factors that relate to the occurrence of multiple brooding in Sebastes spp. 
185 
Table 4 
Summary of predictor variables considered in generalized linear models used to evaluate the 
probability of a rockfish species ( Sebastes spp.) being a multiple brooder. Predictor variables 
include latitude (decimal degrees), temperature (°C), dissolved oxygen (DO) concentration 
(mL/L), depth (m), maximum fork length (L„, cm), maximum age (Aj^^, years), and natural 
mortality rate (M). Relative importance (RI), evidence ratio (ER), log evidence ratio (LER), and 
direction of effect (Direction) on the probability of a species being a multiple brooder for each 
predictor are shown. 
Predictor variable 
RI 
ER 
LER 
Direction 
Latitude 
0.81 
4.27 
0.63 
_ 
DO 
0.64 
1.79 
0.25 
- 
L„ 
0.18 
0.21 
-0.67 
+ 
Temperature 
0.15 
0.17 
-0.76 
+ 
^MAX 
0.07 
0.08 
-1.10 
- 
M 
0.03 
0.03 
-1.58 
+ 
Depth 
0.01 
0.01 
-2.19 
+ 
evolutionary advantage in producing multiple broods 
over those species constrained to a single brood. The 
presumed mechanism for such an advantage would be 
to increase the probability of post-parturition larvae to 
encounter appropriate foraging conditions in this more 
variable, less seasonal region of the California Current. 
This is consistent with the notion of Cushing’s match- 
mismatch hypothesis in which the effect of highly 
variable prey abundance for early life history stages 
of marine fish species may be partially mitigated by 
spreading spawning effort across a broader time period 
(Cushing, 1990). 
Temperature affects individuals directly by influencing 
their growth and metabolism (Boehlert and Yoklavich, 
1983; Boehlert et al., 1991). Temperature also influences 
egg and larval development; for example, recent laboratory 
studies have shown that multiple-brooding rosy rockfish 
( S. rosaceus) produced and released broods more quickly in 
warm water (relative to cool water) treatments, indicating 
faster development of eggs and larvae (Beyer 1 ), and that 
faster development could result in the production of multi¬ 
ple broods during potentially narrow environmental win¬ 
dows for spawning. Temperature may tend to covary with 
other oceanographic variables that drive ocean produc¬ 
tivity and consequently the availability of food resources, 
given that past studies have shown that the condition and 
reproductive output of some rockfish species can degrade 
during the warmer ocean conditions associated with 
El Nino-Southern Oscillation events (VenTresca et al., 
1995; Harvey et al., 2011). 
The importance of depth in this study is consistent with 
other studies that demonstrate that depth is a key driver of 
benthic community structure (Gunderson and Vetter, 2006; 
Tolimieri and Anderson, 2010). We did not include rockfish 
1 Beyer, S. 2018. Unpubl. data. Dep. Ecol. Evol. Biol., Univ. Calif. 
Santa Cruz., 130 McAllister Way, Santa Cruz, CA 95060. 
species of the continental slope in this analysis because of 
our desire to focus on the multiple-brooding phenomenon, 
which is largely limited to species found on the continental 
shelf of the West Coast. However, we recognize that depth 
may more generally drive a broader range of reproductive 
strategies for rockfishes, such that species in deeper waters 
(>400 m) may respond to variable conditions by abortive 
maturity, skipped spawning, or atresia (Nichol and Pik- 
itch, 1994; Lefebvre and Field, 2015; Conrath, 2017), rather 
than the production (or lack thereof) of multiple broods. 
Although we did not formally evaluate the phylogeny of 
rockfish species relative to the capacity to produce multiple 
broods, it would appear that there are examples of both sin¬ 
gle and multiple brooders across the phylogenetic range of 
the genus Sebastes (Hyde and Vetter, 2007). 
The similar predictive ability of temperature and lat¬ 
itude reflects the prominence of the influence of these 
variables on current probability of shelf rockfish spe¬ 
cies being multiple brooders. However, ocean conditions 
are changing. As the effects of climate change become 
more pronounced, the environmental conditions that 
are associated with certain regions (i.e., latitudes) will 
also change. Additionally, we did not find any evidence 
indicating that demographic variables are substantially 
associated with the phenomenon of multiple brooding. 
Although the maximum length and maximum age of 
species did not serve as a substantial predictor of mul¬ 
tiple brooding among species, within species the occur¬ 
rence of multiple brooding has been shown to increase 
with both size and age (Beyer et al., 2015; Lefebvre 
et al., 2018). Our results are also relevant to consider¬ 
ation of how observations of multiple brooding within a 
stock or species may vary with future warming or with 
the increasingly variable ocean conditions expected, and 
potentially already being observed, as a result of climate 
change (Rykaczewski et al., 2015; McCabe et al., 2016). 
For example, Lefebvre et al. (2018) found that during 
