8.4 Energy Metabolism and Provisioning Capacity
Davidson (1997; see also Kay 2004) found that tropical rainforest canopy ecosystems are dominated in both numbers and biomass by several hardworking ant species. These "high-tempo" species all feed on plant or homop-
BOX 8.2 Provisioning and Spatial Patterns of Resource Exploitation
Provisioning mason bees (Osmia lignaria) feed themselves nectar, but deliver both pollen and nectar to their nests. Williams and Tepedino (2003) placed nest boxes so that bees could fly in one direction to a patch of flowers with a high nectar content, but little pollen (Heterophyllum capitatum), or in the opposite direction to a patch with high-pollen but low-nectar flowers (Salix spp.). By examining pollen loads on returning bees, they concluded that the bees must have visited patches ofboth types offlowers on virtually every provisioning trip, in spite of the fact that this must have involved a lot of extra travel. How could this happen?
To model this situation, consider a provisioner with a central place located midway between two food patches. Patch 1 provides food for self-feeding and for delivery to the central place, while patch 2 provides food only for self-feeding. Are there any conditions under which a delivery-maximizing provisioner should visit both patches?
Each time the provisioner visits patch 1, it collects a load of size L for delivery, which requires time tp. Round-trip travel to either patch requires time tt. On each visit to patch 1, the provisioner must spend time ts1 self-feeding to cover the energetic costs of the trip. Alternatively, the provisioner could cover its costs by traveling occasionally to patch 2, where it can achieve a higher self-feeding rate, and spending time ts2 self-feeding. To balance its energy budget using this two-patch tactic, the provisioner must make x (x < 1) trips to patch 2 (on average) for each trip to patch 1.
When using only patch 1, the provisioner achieves a delivery rate of J1 = L/(t t + ts1 + tp). When it uses both patches, it can deliver at rate d2 — L/[(1 + x)tt + x (ts1 + tp)]. The provisioner should use both patches if d2 > d1. Substituting our delivery rate expressions and simplifying gives x(tt + ts2< ts1
In words, it makes sense to use both patches when the time that the provisioner must commit to self-feeding from patch 2 (which includes the travel time there and back) is less than the self-feeding time at patch 1. One can easily find combinations of parameter values for which this condition holds. Although they do not use this modeling format, Williams and Tepedino (2003) propose essentially this explanation for the puzzling provisioning behavior of mason bees.
teran exudates (honeydew)—substances that provide lots of energy but little protein. Workers consume these high-energy foods while they feed protein-rich prey (mostly arthropods) to the brood. The provisioning framework developed in section 8.3 emphasizes the relationship between a provisioner's energy metabolism (expressed as ci, the energy expenditure rate of delivery tactic i), fueled by self-feeding at rate bs —cs, and the consequent delivery rate (called di). These parameters summarize many aspects of an organism's physiological ecology. This section explores some features of energy metabolism relevant to provisioning models and provides a framework for understanding how it is that some species can be so hard-working.
Consider again our stereotypical avian central place forager that works as hard as possible delivering prey to its offspring. Over the past three decades, many studies have sought to characterize energetic capabilities by measuring and comparing energetics in the field. Drent and Daan (1980) argued that free-living birds could not expend more than four times their basal metabolic rate (BMR). Kirkwood (1983) reached a similar value by comparing records of maximum food intake for various species. More recently, Hammond and Diamond (1997) compiled information on the maximum sustained metabolic scope (SMS), measured as the ratio ofmass-adjusted sustainable metabolic rate (SusMR; kJ/gd) to resting metabolic rate (RMR; kJ/gd), and found that it ranges widely among species, but has a median value ofabout four. Similarly, there is great interspecific variation among social insect species in the speed and energetic cost of worker task performance (called "tempo" by Oster and Wilson 1978).
In the most comprehensive review of the avian literature, Williams and Vezina (2001) listed more than fifty field studies that measured the energetic expenditure ofbirds during reproduction, and concluded that our understanding of intraspecific variation remains rudimentary. The fact that two congeneric seabirds, the masked booby (Sula sula) and the gannet (Sula bassanus), show strikingly different work rates (1.6 BMR for boobies and 6.6 BMR for gannets) illustrates the depth of our ignorance. How does this variation arise? Why don't all provisioners work hard? Ecologists usually turn to life history theory to explain this variation, but this subsection considers purely energetic possibilities.
The development of equation (8.2) assumed no limit on the amount of energy that a provisioner can expend, but some sort of limit probably exists. Several things might limit daily energy expenditure, including access to food, the rate at which muscles can generate power to do work, or the rate at which fuel or oxygen can be assimilated or distributed to the musculature (Hammond and Diamond 1997). These limitations could influence provisioners in several ways. If access to food limited the rate of energy expenditure, then a provisioner's time would all be occupied either by self-feeding or by delivering prey. When food is abundant, in contrast, self-feeding would allow a high rate of energy intake, but limitations on muscular activity or assimilation might limit the provisioner's energetic output. In the face of these processing limitations, a provisioner could either spend some ofthe day resting in order to avoid exceeding the maximum sustainable expenditure (as in some seabirds; Houston et al. 1996) or (if assimilation limits the conversion of input to energetic output) use reserves to increase the total daily energetic expenditure.
If the objective is to maximize total daily delivery, a time-limited provisioner should maximize delivery per unit time, and an observer would record performance matching the predictions of a rate-maximizing currency. However, when energy expenditure limits total daily delivery, a rate-maximizing provisioner would reach the expenditure ceiling before the end of the available time, and would have to spend the remainder of its time resting. In this scenario, it makes sense to maximize delivery per unit energy expenditure (i.e., maximize efficiency), and models show that less expensive options that deliver at a lower rate can achieve a higher overall rate of delivery in these situations. In this case, an observer would record behavior that matches the predictions of efficiency or modified efficiency currencies. These issues are treated more fully by Hedenstrom and Alerstam (1995), Ydenberg (1998), Houston and McNamara (1999), and Nolet and Klaassen (2005). Thus, provisioning behavior operates within an envelope bounded at one extreme by rate maximization and bounded at the other extreme by efficiency maximization. Self-feeding rates determine the predicted behavior: high self-feeding rates should shift provisioners toward higher workloads, while low self-feeding should have the opposite effect.
Why should some provisioners have high energy capacities while others have low energy capacities? Hammond and Diamond (1997) suggest that animals with high energy capacities need expensive metabolic machinery, including organs with high metabolic rates such as the liver, heart, and kidneys (Daan et al. 1990; see also section 5.3). Enhanced metabolic performance (e.g., Suarez 1996, 1998) can evolve under selection or develop on a physiological time scale within individuals, but it always comes at the cost of a metabolic machine that is more expensive to run. So why does it make sense for some animals to maintain this expensive machinery while others do not?
In the framework developed here, potential delivery depends not only on prey availability but also on the provisioner's capacity for hard work, which in turn requires fuel from self-feeding. If high delivery rates enhance fitness, then better self-feeding opportunities also allow increased metabolic capacity. For example, dominant individuals probably have better access to food (e.g., Hogstad 1988), and so can support higher metabolic rates. Bryant and Newton
Daily energy expenditure Daily self-feeding
Figure 8.5. Hypothetical scheme of relationships between total daily delivery (on y-axis), resting metabolic rate (r), self-feeding rate, daily energy expenditure, and maximum daily energy expenditure (k) (all onx-axis). Total daily self-feeding and total daily energy expenditure must balance. As self-feeding increases, the provisioner can expend and deliver more. The lower curve shows the relationship for a "down-regulated" metabolism, and the upper curve for an "up-regulated" metabolism. Up-regulation gives higher deliver capacity, but also generates a higher resting metabolic rate, and so is more expensive to maintain. In the example shown, the provisioner benefits from up-regulated metabolism when the attainable self-feeding rate exceeds the rate labeled s*.
(1994; see also Hogstad 1987) interpret the higher BMR of dippers (Cinclus cinclus) as a cost of dominance, but if high BMR translates into better provisioning, the chain of causation could be reversed: the higher BMR may be a benefit of dominance. Figure 8.5 shows these relationships diagrammatically.
Davidson (1997) suggests that some ant species dominate rainforest canopies because they can easily obtain high-energy exudates that favor the evolution of high tempos. These high-tempo ant species can deliver protein resources at a high rate and can vigorously defend territories, traits that in turn lead to the high reproductive rates that give these species their dominance. This example shows the broader ecological consequences of provisioning behavior.
Other social insect studies suggest similar relationships. Four honeybee species in the genus Apis show marked differences in worker mass-specific metabolic rate, colony metabolism, and the intensity of provisioning. The hightempo species (A. mellifera and A. cerana) deliver more resources than the low-tempo species (A. dorsata and A.florea) and produce offspring at a greater rate, but have higher worker mortality (Dyer and Seeley 1991). We can examine the great interspecific variation in metabolic expenditure documented among birds in the same way. In fact, some investigators hypothesize that endothermy itself—a particularly expensive metabolic mechanism—has evolved to enhance aerobic capacity and support the vigorous exercise required in many forms of provisioning and parental care (Farmer 2000).
As these examples imply, close ecological relationships connect the energetically expensive activities of animals (such as provisioning), the foraging behavior that obtains the fuel, and the metabolism that powers the activity. Nonetheless, few studies have as yet considered these ideas and their implications.
An increase in brood size or an approaching period of shortage can increase the demand for delivered food. None of the provisioning models described so far consider variations in demand. Their assumptions and equations incorporate only "supply-side" parameters such as encounter rate (e.g., waiting time), travel distance, self-feeding rate, and the energetic cost of delivery. This section considers how demand may affect provisioning. What properties should we expect provisioning strategies to have for responding to demand?
A large number of excellent studies on birds and social insects demonstrate that provisioners respond to natural variation in demand, increasing the delivery of materials when demand rises and vice versa. We expect this, of course; demand varies in nature. Demand can vary predictably (e.g., nestlings grow), or unpredictably (e.g., bad weather). Experimentalists have used a variety of methods to manipulate demand. In studies ofbirds, the most common experimental technique manipulates clutch or brood size. Interest has been concentrated on the fitness consequences for parents and offspring, with little attention given to the tactics parents use to increase delivery. Social insect studies have usually focused on how colonies recover their populations and stored reserves after an experimentally imposed demand. A few studies have considered the behavior of individual workers (Cartar 1991; Schmid-Hempel et al. 1993).
Provisioners might use several basic tactics to meet increased demand. First, provisioners could simply spend more time provisioning. The longer hours of work will necessarily increase their daily energy expenditure. Second, a provisioner can work harder—for example, by flying faster. When experimenters remove pollen stores from a beehive, more workers collect pollen, and each individual worker works harder at the task (Eckert et al. 1994). Third, provisioners can use energy from body reserves to fuel extra delivery effort. And finally, provisioners can alter the selection of prey for delivery. I discuss fueling of delivery from reserves and prey selection in the next few paragraphs.
Provisioners can overcome the limits imposed by self-feeding on energetic expenditure by using reserves to fuel activity (Houston 1993). Many studies have reported reductions in the body mass of parent birds during periods of intense provisioning activity (e.g., Moreno 1989). Most researchers take this to indicate that energy expenditures exceed energy intake during provisioning. A parallel body of results describes the responses of social insect colonies, especially honeybees, to the challenges of imposed parasite loads (Janmaat et al. 2000) or removal of pollen (Fewell and Winston 1992). These studies commonly show that colonies reduce their reserves as they react to the manipulation and recover to their former state.
In analyzing whether provisioners should "dip into" their reserves to address an unexpected demand, most models assume that reduction ofthe provi-sioner's stores has life history costs, such as an increased risk of starvation. The provisioner must balance these costs against the advantages ofincreased delivery capacity. Nur's (1987) model provides a paradigmatic example. It seeks to explain patterns in the provisioning responses of songbird parents to (manipulations of) brood size. The model assumes that increased feeding frequency increases nestling weight and survival, but reduces parental weight and survival. Nur concludes that parents feed larger broods more frequently because the greater fitness value of the brood makes increased effort worthwhile (see also Beauchamp et al. 1991).
Swifts (Apus apus) delivered more food to experimentally enlarged broods, but each nestling received less, resulting in a lower mean chick mass (at age 12 days; Martins and Wright 1993b). In addition, Martins and Wright found that parents lost mass during the provisioning period, and they assumed that a reduction in self-feeding caused this weight loss. They argued that this weight loss imposes a risk on parental survival, but others have suggested that provisioning swifts may lower their body weights to reduce flight costs. Neither idea seems complete. Why is mass loss risky, given that parents can recover quickly? On the other hand, if lowering mass lowers flight costs, why don't all parents lose weight?
Thinking about the relationship between the use of a fat reserve and self-feeding suggests a slightly different explanation. A small reserve of fat could boost a parent's delivery rate, either by providing the power for a period of energetic expenditure above the sustainable limit or by reducing the need for self-feeding for a brief period. A provisioner could expend this fat reserve slowly and steadily on a programmed schedule to meet a predictable increase in demand (e.g., growing nestlings), or it could expend its reserve all at once to meet demand during unpredictably poor conditions (e.g., cold, rain). Under many circumstances, the provisioner would be able to "restock" its reserve when conditions improve. The key idea here is that expending the reserve need not endanger the provisioner's survival. The provisioner maintains the reserve (presumably at a small ongoing cost to provisioning capacity) in order to buffer provisioning capacity against fluctuations in prey availability.
Was this article helpful?