Unformatted text preview: ANT 154BN Course notes Lecture #5: Feeding ecology II and foraging theory 18 January 2011
Key terms and concepts are indicated in blue Outline
1. Feeding selectively (continued) 2. Optimal foraging theory 1. Feeding selectively (continued)
Primate nutritional ecology (principles from last time, based on Leighton 1993) • Many factors potentially influence food choice • Many of these factors co-vary (i.e., protein, calories, fiber) • Multivariate analyses are required to identify the importance of different components on selectivity • Multivariate models should include all potentially important variables Example #3: Bornean orangutans (Leighton 1993, the reading for today) Some improvements over previous two case studies: • used independent, unbiased samples of feeding behavior (proportion of independent feeding obs) • incorporated full range of ecological, morphological, and chemical parameters • empirically measured selectivity (relative use/relative availability)
272 Fruiting phenology and dietary intake variables
~k 70 60 50 40 30 20 10 0 Food tree diversity
i i i Leighton (a) /%.
12 1a 6 20 ' 2'4 4 24 8 Food tree density 20 16 12 8 (b) 42
| i i i i i 4 8 12 16 20 24 2 1.41.2' 1.0 0.8 i Fig availability
k= e, ~
i , . 9 i 9 i . I c) 0.60 . 4 -'
0.2" 0.0 0
i 4 8 12 16 2O 24 1.0"~
o o .- ~
0.8- (d) Dietary composition 0.6- o 0 0 ~t. .E
0 "0 0.40.2- 9 bark+leaves pulp+seeds 0.00 4 8 12 16 20 24 m o n t h of s t u d y Fig. 1. P h e n o l o g y o f f r u i t p r o d u c t i o n a n d o f o r a n g u t a n d i e t d u r i n g 11 c o n s e c u t i v e i n t e r v a l s (see T a b l e I) o f t h e 2 4 - m o n t h study ( m o n t h 1 = S e p t e m b e r 1977). (a) Diversity ( n u m b e r o f species) o f large t r e e s ( > 2 5 - c m d b h ) r i p e n i n g fruit 9 (b) D e n s i t y o f large t r e e s a t p e a k r i p e f r u i t p r o d u c t i o n . ' (c) D e n s i t y o f r i p e fig crops o f large h e m i e p i p h y t i c a n d c l i m b i n g fig p l a n t s . ( d ) S e a s o n a l c h a n g e s i n p r o p o r t i o n s o f o b s e r v a t i o n s o f o r a n g u t a n f e e d i n g o n t h r e e types o f p l a n t food. Leighton 1993
7 Tuesday, January 19, 2010 ANT 154B Lecture #5 course notes
Figure 2 Fruiting behaviour over 68 months (b) (f) page 2 of 6 in a Bornean rainforest for different plant Figs are always available and forms. (a) Figs: average number of stems observed each month = 116 (min = 93, max = 121); (b) Woody climbers: average Landscape level Bornean plant reproduction 961 Landscape level Bornean plant reproduction 961 Landscape level Bornean plant reproduction 961 number of stems observed each month = 951 (min = 847, max = 991); (c) (c) (a) (e) Small trees (14.5–24.5 cm DBH); (d) Large (a) (a) (e) (e) trees (> 24.5 cm DBH). Observed values are shown in thick grey line. The average level of fruiting expected across all months is indicated by the solid black line while 95% conﬁdence limits are shown by the dashed black lines. The thin grey line illustrates a single replicate of random fruiting behaviour. (d) (f) (e) Frequency (b) distribution of ﬁg reproductive (f) (f) (b) (b) levels by month. (f) Frequency distribution of woody climber reproductive levels by iting behaviour over 68 months (g) Frequency distribution of small month. iour over 68 months plant rainforest for different rainforest for different plant tree reproductive levels by month. (h) Freigs: average number of stems t for different plant igs: average number of stems quency ch month = 116 (min = 93,distribution of small tree reproducch number = 116 (min = 93, e month of stems tive levels by month. Barcharts illustrate (b) 116 (min = 93, average (b) Woody climbers: average = Woody climbers: f stems observed stems each observed levels of reproduction. Black fy climbers: observed average each (min = 847, max = 991); (c) 991); assume a single season, grey curves (min = 847, max =curves(c) (c) (g) (c) (g) observedDBH); (d) Large each 14.5–24.5 cm 14.5–24.5 cm DBH); assume a mixed model with two seasons. (d) Large 47, max = 991); (c) Fig phenology unrelatedtree phenology phenology to tree their phenology is unrelated to
(g) Figs are always available Lianas (h) Figs cm DBH);The average level of k grey line. (d) Large ted across all months is indica. ted across all months is indicaObserved values are Cannon, Curran, Marshall & Leighton (2007) Ecol. Lett. solid black line while 95% while . solid averageline Tuesday,95% e 19, 2010 The black level of January and alluvial bench, two Ômid-seasonÕ months are between taxonomic diversity and increasing reproductive sandston 8 mitsmonths is indica- dashed mits are shown by the dashed all are shown by the levels, primarily the lowland sandstone and peat swamp. apparent. he thin while illustrates a This pattern suggests that the Ômid-seasonÕ period ck line grey line95% te of random fruiting behaviour. This patt and seeds may largely (d) due to slight differences in the timing of be te of random fruiting behaviour. (h) hown by the ﬁg reproductive Reports a general “preference” for fruit pulpern was most pronounced in the peat swamp, (d) (h) distribution of dashed distribution of ﬁg reproductive where 2almost all months with > 5% reproduction fellLeighton 74 rey (f) Frequency distribution line illustrates reproduction among the forest types. Lowland granite has a nth. nth. (f) Frequency distribution signiﬁcantly below expectations (Fig. 5g). Phylogenetic m fruiting behaviour. loby limber reproductive the west AF level (slightly > 2%), while SMF levels (12%) levels 16 (a) (d) (h)requency distributionwere am of n of ﬁg distribution of small diversity ==in each forest type was consistently below requencyreproductive small ong the highest (Fig. 4c,j). Marshall, Boyko, Feilen, Boyko, & Leighton (20 + .= ctive levels by month. (h) Freequency by month. Both ctive levelsdistribution(h) Fre- in absolute terms and in the correlation with t3L expectations, exc2ept in the montane and lowland granite 1. ~ bution of small tree reproducbution of small tree by Tuesday, January 19, 2010phylogenetic diversity with increasing roductive levels reproduc(Fig. Food tree == 1 . 0 y month. Barcharts reproduction levels, forest types differed signiﬁcantly in the density 5h,j). Declines in y month. Barcharts illustrate illustrate distribution of small ~ =,. vels of reproduction. Black of reproduction. reproductive5 participation were only weakly apparent in patterns o vels month. (h) Fre- Black of taxonomic and phylogenetic diversity relative to by single season, grey curves o ea e a single season, grey curves r, most forest types (Fig. 5). reproductive levels (Fig. 5). Montane forest had the lowest small tree reproduced model with two seasons. = "6 ed model with two seasons. of expected taxonomic diversity (Fig. 5a) while the y ss 2 + 0.96x r2 = 0.8 The correlation of reproductive levels acro= 0.3forest (types5) levels Barcharts illustrate i i o 012 018 1.0 112 was quitet~ strong o when the0 4 SMF016events were included 1'4 lowland sandstone had the highest (Fig. 5d). The general eproduction. Black season, grey curves Ômid-selevel months are (Table ductive expected asonÕ of phylogenetic diversittaxonomic ilar rsity and increasing repro1). Reproduction acrof ssi t iall plants> 2in cthed b h )reea welly was simdive across all l o g r u n g t r e e s ( 5 m th / h between nd alluvial bench, two between taxonomic diversity and increasing reproductive nd alluvial bench, two Ômid-seasonÕ months are with two seasons. that th Ômid-seasonÕ period drained lowland "(0 levels, primarily the lowland sandston his pattern suggests that the Ômid-seasonÕ period levels, primarily the lowland sandstone and peat swamp. his pattern suggestsforest etypes, indicating that while species numbers may be e and peat swamp. 1.6- forest types (lowland granite, lowland 9 (n This taxa come from a diverse be due to slight differences in the timing of + sandstone and lower in in montane of This pattern was most pronounced in the peat swamp, 1.4be due to slight differencesthe the timing forest, thesepattern was most pronounced in the peat swamp, alluvial bench) were strongly correlated while r where almost all month e trends n among the forest types. Lowland granite has types. lineages. Few forest types exhibited negativs with > 5% reproductionr fell 1.2~ where almost all months with > 5% reproduction fell n among the forestset of Lowland granite has the uppe and lowland granite forests were also tightly signiﬁcantly below expectations increasing reproductive F level (slightly > 2%), while SMF levels (12%) level (slightl > 2%), while months are signiﬁcantly below expectations (Fig. 5g). Phylogenetic lFbench, twoyÔmid-seasonÕ SMF levels (12%) between taxonomic diversity and (Fig. 5g). Phylogenetic"o 1.0" diversity in each forest type was consistently below 0.8" the highest (Fig. 4c,j). diversity in each forest type was consistently below the highestthat the Ômid-seasonÕ period levels, primarilyFood tree sandstone and peat swamp. the lowland diversity suggests (Fig. 4c,j). 8 expectations, except in the montane and lowland granite 0.6absolute terms and in the correlation with Ó 2007 Blackwell Publishing Ltd/CNRS expectations,was ept exc most pronounce in the peat swamp, granite absolute termsrencesin the correlation tolevels, forest types differed signiﬁcantlyofwith This patternDeclines inin the montanedandtylowlandcreasing"6 slight diffe and in the timing in the (Fig. 5h,j). phylogenetic diversi with in n 0.4-~ ~ 1 + 1.07x a(r2 = 0.94) the forest types. Low tic diversity relative reproductive all month with > wea reproduction 0,2 taxonomic and phylogeneland granite has to where almost participations were only 5% kly apparent in fell 0.6 0.8 ' 1'.0 1'.2 1'.4 1i 6 1'.8 most forest types (Fig. 5). elightly (Fig. 5). Montane forest had (12%) levels > 2%), while SMF levels the lowest signiﬁcantly below expectations (Fig. 5g). Phylogenetic log f r u i t i n g t r e e d i v e r s i t y (no. s p p . ) The in ation forest type was consi forest types pected taxonom est (Fig. 4c,j). ic diversity (Fig. 5a) while the diversity correleach of reproductive levels across stently below The correlation of reproductive levels across forest types pected taxonomic diversity (Fig. 5a) while the was quite except in the SMF events lowland granit dstone had the highest (Fig. 5d). The general terms and in the (Fig. 5d). The genera was quite strong when the SMF events were included f i dstone had the highest correlation with l expectations, strong whenthe montane andwere included 0 .e(c) 10 (Table 1). Reproduction across all plants in the three creasing el of types differed signiﬁcantly in the (Table 1). Reproduction across all plants in the three wellel of phylogenetic diversity was similar across all orestphylogenetic diversity was similar across all (Fig. 5h,j). Declines in phylogenetic diversity with inwell- 0.5r drained lowland forest types (lowland granit lowland , indicating that while species numbers may be species numbers to drained lowland forest types (lowland granite, lowland 0 . 4 , indicating that whilediversity relative may be reproductive participation were only weaklye, apparent inc and phylogenetic ;= Figs, bark leaves are fallback foods for while 0 . 3 sandstone and alluvial bench) were strongly correlated while montane forest, these taxa come from a diverse come from a diverse montane forest, these taxahad the lowest and sandstone and alluvial bench) were strongly correlatedorangutans (cf gibbons) most forest types (Fig. 5). ig. 5). Montane forest the upper and lowland granite forests were also tighytly ges. Few forest types exhibited negative trends Modeling tigh 6 27Leighton 1993 5 the upper and lowland granite forests were also D i e t a r "ttly S e l e c t i v i t y b y Orangutans ges. Few forest types exhibited negative trends - cm DBH). Observed values are(c) cm DBH). Observed values are (g) General “preference” for fruit pulp & seeds
1.4- 0.80.6- 0.40.2 eo o Q. EIL C Figs and bark/leaves are fallback foods
m B B xonomic diversity (Fig. 5a) while the ad the highest (Fig. 5d). The general logenetic diversity was similar across all g that while species numbers may be forest, these taxa come from a diverse forest types exhibited negative trends The correlation of reproductive levels across forestOtypes20, Tuesday, January SMF was quite strong when the 19, 2010events were included e} Q. 0 . 60,1 (a) 0 Ó 2007 Blackwell Publishing Ltd/CNRS Ó 2007 Blackwell Publishing Or (Table 1). Reproduction across all plants in theLtd/CNRS 0 . 5 -0lthree wel .0 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 drained lowland forest types (lowland granite, lowland 0.4f r u i t i n g figs / 10 ha "6 sandstone and alluvial bench) were strongly correlated while F i g . 2 . P r e f0e3r-e n c e f o r p l a n t f o o d t y p e s a s i n f e r r e d f r o m t h e c o r r e l a t i o n b e t w e e n t h e r e l a t i v e F granit the upper and lowlandigs e forests were alsoo8,r ttightlyb s e r v a t i o n s o f o r a n g u t a n f e e d i n g o n e a c h t y p e a n d its a v a i l a b i l i t y d u r i n g 11 pr po ion of o 10 O Ó 2007 Blackwell Publishing Ltd/CNRS o. 0.0 cl s a m p - e p e r i0o2 -s ( T a b l e I ) . ( a ) C o m b i n e d s e e d - a n d f r u i t p u l p - f e e d i n g w a s d i r e c t l y r e l a t e d t o .d o t h e E e n s i t y o f large fruit c r o p s ( o f large trees). ( b ) T h e relative f r e q u e n c y o f fig-eating was d u n r e l a t e d t o . 1h e a v a i l a b i l i t y o f r i p e fig f r u i t p a t c h e s , i n d i c a t i n g t h a t t h e y w e r e n o t p r e f e r r e d . 0 t 9 i 0.4 9 i 09 9 | 0.6 9 i 0.7 of all B ! 0.8 feeding B j 0.9 bouts , I~ 0.3 1.0 proportion (pulp+seeds) 0.56} > "
0.4" (b) 0.3- Bark/leaves 0.2O O o. O O. 0.1 D E! Tim Laman
Q 0.0 03 i 0.4 proportion i 0.5 i 0.6 i 0.7 of all i 0.8 feeding i l:! , 0.9 bouts 1.0 (pulp+seeds) Fig. 3. T h e p r o p o r t i o n s o f o b s e r v a t i o n s o f ( a ) f i g - a n d o f ( b ) leaf+bark-feeding by orangutans were both inversely related to combined fruit pulp+seed-eating. Leighton 1993
t e n d e d t o b e e a t e n d u r i n g t h e s a m e p e r i o d s (rs = 0.68, P < 0.05). T h e r e Tuesday, January 19, 2010r e , c o n s u m p t i o n o f t h e p r e f e r r e d types, n o n f i g r i p e p u l p a n d u n r i p e s e e d s , fo w a s l i m i t e d b y availability; t h e s e f o o d s c o m p r i s e d > 8 5 - 9 0 % o f all f e e d i n g o b s e r v a t i o n s w h e n f r u i t i n g t r e e s e x c e e d e d 4 t r e e s / h a (Fig. 2a, T a b l e I).
11 ANT 154B Lecture #5 course notes Factors influencing selectivity: Multivariate analyses show selection for: 1. large crops 2. high pulp weight/seed 3. high pulp weight/fruit page 3 of 6 Orangutan fig selection: incorporating chemical components Select for: 1. large crops 2. high pulp weight/seed 3. high digestible carbohydrates 4. low phenolics 2. Optimal foraging theory
General principles: Optimality theory • well developed theory • results (generally) robustly supported • mathematically specifies costs and benefits • used in a variety of contexts Optimality models: foraging currency = net rate of energy uptake (E/T) where: E = energy gained (calories) T = time expended Assume: animals evolve to maximize E/T Diet breadth model (aka, prey choice model, contingency theory): Predicts solutions to a basic question: to feed or not to feed (on a particular item)? Assumptions: 1. Selection acts to maximize long-term rates of energy gain (E/T). 2. Feeding and searching are mutually exclusive. 3. Prey encountered sequentially (i.e., not simultaneously). 4. Within a prey type, energetic value of prey, handling time, & search time constant. 5. Animal has “complete information” ANT 154B Lecture #5 course notes Terms: Ei : the energy obtained from from item i hi : the handling costs of processing item i si : the search costs of finding item i (related to encounter rate) Profitability of a food item Diet breadth is Ei/hi page 4 of 6 model New item i is adde d to the diet as long as: Ei / hi ! E / ( s + h) where: E : average energy content of fo o d items s : average search time for fo o d items h : average han dling time for fo o d items Diet breadth model Tuesday, January 19, 2010 26 Ei / hi ? E / ( s + h) the new item un der consi deration the average diet (a depiction of a particular place an d time) Tuesday, January 19, 2010 27 Diet breadth model: implications 1. If average search times are long and handling times are short, then forager should take most prey that it sees . (i.e., generalize) 2. If search times are short and handling times long (s is small and h is big), then forager should take only high E/h items. (i.e., specialize) 3. For unprofitable items, abundance is irrelevant. 4. Foragers should generalize when environment unproductive, and specialize when the environment is productive ANT 154B Lecture #5 course notes Marginal value theorem: Predicts solutions to another basic question: how long to feed in a food patch? page 5 of 6 Assumptions: 1. Selection acts to maximize long-term rates of energy gain (E/T). 2. Resources are patchy on spatial scale of forager movements. 3. Feeding and searching are mutually exclusive. 4. Prey encountered sequentially (i.e., not simultaneously). 5. Animal behaves as though it has “complete information” 6. Patches differ in their quality (i.e., E/T) Marginal value theorem: core model
cumulative calories consumed searching time
high handling and eating time high travel time 0 time departed previous patch
Tuesday, January 19, 2010 patch residence time giving-up time high Charnov 1976
39 Effects patch density: when more dense patches, spend less time per patch Effects of patch quality: when higher quality patches, patch residence time decreases Marginal value theorem: implications 1. Leave all patches when rate of gain equals that found at point where the line drawn from point of departure from last patch is tangent to the energy gain curve. 2. Stay less long in patches when travel time between patches is short (i.e., when density of patches is high). 3. Stay less long in patches when overall patch productivity (i.e., patch quality) is high. ANT 154B Lecture #5 course notes Other considerations: page 6 of 6 1. Diet breadth and marginal value models are based on mean rates of energy intake. Variance (i.e., riskiness) also important. 2. Animal’s information about environment not always perfect. 3. Energy (i.e., calories) may not be the only important factor (e.g., micronutrients, minerals). 4. More efficient travelers should leave patches more quickly (because travel times reduced). 5. Other factors (e.g., predation risk, mating and parenting strategies) also affect foraging decisions. Take home messages
1. Orangutans appear to select predominantly based on patch size and caloric returns, and incorporate a number of other chemical and nutritional elements into foraging decisions. 2. Optimal foraging theory assumes that animals forage in such a way as to maximize E/T. 3. The diet breadth model explains why many potential food items are not consumed, and predicts different feeding strategies under different environmental conditions. 4. According to the MVT, patch residence times are determined by patch quality and the density of food patches.. Question to ponder
The foraging models presented today are based on a number of underlying assumptions. Discuss the extent to which these assumptions seem appropriate (e.g., which seem reasonable, which are most likely to be violated by foraging primates?). How might the unusually gregarious nature of most primate species influence the extent to which these basic assumptions are valid? ...
View Full Document
- Spring '10
- Forest, Optimal foraging theory, Leighton, forest types