Desynchronization and re-synchronization of reproduction by Astragalus scaphoides, a plant that flowers in alternate years

0 E. E. Crone (&) Harvard Forest, Harvard University , 324 N. Main Street, Petersham, MA 01366, USA Mast seeding, the synchronous seed production by plants at irregular intervals, has been widely studied from the perspective of its fitness benefits, but much less is known about the proximate factors that cause plants to reproduce synchronously. In this article, I follow up on more than two decades of research investigating proximate mechanisms of mast seeding by Astragalus scaphoides, an iteroparous perennial forb. We use longterm monitoring in relation to two environmental manipulations to evaluate the importance of exogenous environmental factors versus endogenous feedbacks for synchrony in this species. Our past research showed that synchrony in this species is explained by the pollencoupling hypothesis: plants that flower synchronously set seed and deplete stored resources, whereas plants that happen to flower asynchronously are pollen limited, set fewer seeds, and do not deplete resources, and flower again until they are resynchronized. Continued monitoring of two experimental manipulations, water addition, and flower removal, provides additional support for this model, and also reveals subtle effects of water availability on synchrony. Water addition decreased flowering, rather than increasing it as expected based on simple correlations with weather variables, suggesting that precipitation does not synchronize reproduction. Exogenous drivers are generally considered to be the primary synchronizing factors in plant reproduction. Our work in this system suggests that endogenous feedbacks may be more important than has been previously assumed. - New insights into mechanism and evolution of mast flowering: feedback between theory and experiment Population-level fluctuations of seed output, known as mast seeding or masting, are a relatively common phenomenon in plants. Masting has the potential to have far-reaching effects on community and ecosystem dynamics (Janzen 1976; Ostfeld and Keesing 2000; Kelly and Sork 2002). A large body of literature has addressed the reasons why masting occurs from an evolutionary perspective, in other words, fitness advantages of synchronous reproduction (reviewed by Kelly and Sork 2002). Fewer studies have looked at how masting occurs from a proximate ecological perspective. In general, we do not know how trees are able to synchronize reproduction at super-annual time scales. Nonetheless, understanding how synchrony arises is central to predicting how changes in the environment would affect patterns of seed production by plants, and subsequent dynamics of seed consumers and plant communities. Investigations of proximate mechanisms of synchronous mast seeding have drawn on two categories of explanations, which are not mutually exclusive. The first category explains mast seeding based on exogenous environmental factors, i.e., synchronous environmental forcing of individual reproduction through fluctuations in resource availability (Norton and Kelly 1988; McKone et al. 1998; Schauber et al. 2002; Abrahamson and Layne 2003; Kelly et al. 2008), or cues for flower induction (Ashton et al. 1988; Piovesan and Adams 2001), also known as a Moran effect (Liebhold et al. 2004a). The second category invokes endogenous mechanisms of synchrony, i.e., feedbacks within or among individuals that lead to synchronous dynamics through time, also known as phase locking (Liebhold et al. 2004a). Isagi et al. (1997) and Satake and Iwasa (2000, 2002) formalized this mechanism in the context of mast seeding. These resource-budget and pollencoupling models assume that an individual plant requires more resources to flower and set seed than it gains in a year, and therefore flowers only above some threshold amount of stored resources. These rules can cause plants to have cyclical or chaotic patterns of reproduction over time (Isagi et al. 1997). In the presence of cyclical reproduction by individuals, only small amounts of environmental variation (e.g., frost events that kill buds) are needed to synchronize individuals within plant populations (Satake and Iwasa 2002). In addition, if plants are pollen-limited in low-flowering years, synchronous mast seeding could occur even in the absence of any environmental variation (Satake and Iwasa 2000). Exogenous mechanisms remain the most common explanation for synchronous mast seeding (Koenig and Knops 1998; Schauber et al. 2002; Kelly and Sork 2002; Kelly et al. 2008). However, three modeling studies have shown that endogenous resource dynamics are necessary to explain synchrony, whereas exogenous factors alone cannot (Rees et al. 2002; Crone et al. 2005; Lyles et al. 2009). For the past two decades, my colleagues and I have studied causes of synchronous reproduction by Astragalus scaphoides, a perennial herb that flowers in alternate years (Lesica 1995; Crone and Lesica 2004). Initially, we believed that reproduction might be synchronized by exogenous factors, particularly precipitation during bud initiation, in this semi-arid environment. However, we found at best weak effects of environmental drivers on flowering dynamics (Crone and Lesica 2004, 2006; Crone et al. 2005). Instead, our research has shown that flowering is synchronized largely by endogenous processes, not by external cues such as fluctuations in resource availability. When individual plants flower synchronously with other plants in the population, seed set depletes stored nonstructural carbohydrates (NSC) which prevents flowering the following year (Crone et al. 2009). When plants flower asynchronously, they are pollen-limited, and set fewer seeds (Crone and Lesica 2006), which prevents NSC depletion (Crone et al. 2009). Therefore, these individual plants should flower in subsequent years until they become synchronized with others in the population. These mechanisms largely conform to the pollen-coupling model, and variants of these models fit to our data show that endogenous processes are both necessary and sufficient to explain synchronous flowering in this species (Crone et al. 2005). In this paper, I follow up on three of our past analyses to evaluate whether flowering dynamics after our initial publications have continued to support the importance of endogenous processes in synchronizing reproduction in this species. I also use these follow-up studies to test whether longer-term monitoring revealed stronger signals of exogenous environmental drivers. The first of these studies is an analysis of monitoring at three sites (Crone and Lesica 2004; Crone et al. 2005), in which we previously analyzed patterns from 1986 to 1999, and for which monitoring is ongoing. The second is an experiment in which we attempted to desynchronize plants in these plots by adding supplemental water from 2000 to 2002 (Crone and Lesica 2006). The third is an experiment in which we attempted to desynchronize plants by preventing seed set in 2005, a high-flowering year (C (...truncated)