The Condor 94:912-923 0 The Cooper Ornithological Society 1992

A COMPARATIVE ANALYSIS OF THE NUTRITIONAL QUALITY OF MIXED AND EXCLUSIVE DIETS FOR YELLOW-VENTED BULBULS

IDO IZHAKI Department of Biology, Universityof Halfa at Oranim, Tivon, 36910, Israel

Abstract.Yellow-vented Bulbuls (Pycnonotusxanthopygos) were fed a mixed diet of four fruit species(Rhamnus alaternus,Lonicera etrusca,Rubia tenuifolia, Ephedra aphylla) and diets containing each of these four separately. A mixed-fruit diet proved no more nutritious than a single-fruit diet, and all the birds lost weight at the same rate regardless of diet. Birds fed a mixed-fruit diet assimilated less protein than those fed exclusively on Rubia and Rhamnus fruits, and assimilated more energy than those fed exclusively on Ephedrafruits. The apparent metabolizable energy(AME) in bulbuls fed a mixed diet (0.73) was higher than in those fed on Rubia and Ephedra (0.69 and 0.61, respectively)but lower than in those fed on Rhamnus (0.82). Protein content alone does not explain loss of body mass. However, lack of specificamino acids, or high potassiumto sodium ratios, may cause mass loss. In addition, a mixed diet of four fruit speciesmay not prevent the accumulation of secondarycompounds to a damaginglevel resulting in low protein assimilation. Key words: frugivory;fruit preference:nutrition; Yellow-ventedBulbul; Pycnonotusxan- thopygos;digestion; metabolizability.

and Safiiel 1985, 199 1, unpubl. data), mixed date INTRODUCTION palm-citrus orchards in Iraq (Al-Dabbagh et al. The relationship between which produce 1987), gardens and palm groves in western Ara- fleshy fruits and their dispersers is based partially bia (Meinertzhagen 1954) and the Ethiopian on the nutritional value of the fruit for the dis- lowland forest and non-forest habitats (Moreau perser. The ability to digest fruit is believed to 1972). The northern limit of the distribution of be an important constraint on the evolution of the Pycnonotidae in the Old World is in the Mid- the interaction between plants bearing fleshy- dle East which is populated by the Yellow- fruits and fi-ugivorous birds (e.g., Herrera 1984, vented Bulbul Pycnonotus xanthopygos. The bul- Jordan0 1988, Worthington 1989). bul consumes and dispersesseveral fruit species In previous experiments, frugivorous birds of in the Israeli scrublands (Izhaki 1986; Izhaki the eastern Mediterranean avifauna were fed on and Safriell985,1990b, 199 l), and is one of the a single fruit speciesexclusively (Izhaki and Saf- most important frugivorous birds in the Israeli riel 1989). Although birds ate large amounts of avifauna. In this study, I fed these bulbuls four fruit they nevertheless lost mass due to poor ni- fruit speciescommonly available from June to trogen assimilation. It was assumedthat nutrient August in the northern parts of Israel (Rhamnus intake would be better balanced by feeding the alaternus, Lonicera etrusca, Rubia tenuifolia, birds on several fruit species rather than on a Ephedra aphylla, unpubl. data). These foods are single one, since a mixed diet could increase the relatively poor in proteins and lipids, but rich in possibility of extracting specific nutrients (e.g., and water. The profitability of Jordan0 1988, Mack 1990) and could reduce ac- these fruits is reduced by their poor protein and cumulation of specific secondary compounds lipid content as well as their indigestible seeds (Jordan0 1988; Izhaki and Safiiel 1989, 1990a; (Herrera 1981, Levey 1986, Jordan0 1988). Levey and Karasov 1989). In the current paper, I investigate the following The Pycnonotidae are well-known frugivores questions. (1.) Are bulbuls fed on a mixed diet in regions of the Old World such as the Malay- of several fruit speciesbetter able to maintain a sian lowland rain forest (Lambert 1989) scrub constant body mass than bulbuls fed exclusively and bush-jungle in India (Whistler 1949) eastern on single fruit species?(2.) Does a mixed fi-uit- Mediterranean scrubland (Izhaki 1986; Izhaki diet improve protein, , , and energy digestion? and (3.) Is the choice of a ’ Received 17 February 1992.Accepted 10 July 1992. fruit speciesby birds connected with a particular

[9121 NUTRITIONAL QUALITY OF FRUIT DIET 913 nutritional or morphological characteristicof the daily from several different plants within each fruit? species. After the bulbuls’ mass stabilized, three groups METHODS of six different individuals were offered an ex- clusive diet of one fruit specieseach, and a fourth FEEDING TRIALS group of six birds was offered a mixed diet con- I collected fruits and birds for thesetrials around taining the four fruit species.In each experiment the University of Haifa at Oranim campus, Israel I used six birds picked randomly from the pool (32”43’N, 35”07’E), during the summer and fall of 30 bulbuls. Lonicera, however, was too rare of 1989 and 1990. Bulbuls were captured using in the area for an experiment to use this fruit mist netting and held in individual indoor cages exclusively. Each type of food was provided ad (40 x 50 x 80 cm) fitted with easily cleaned libitum for several days, as well as water ad li- undercarriages.The bulbuls were maintained at bitum, at 08:OOhours. In the trials using all four room temperature (25” * 3°C) with artificial light fruit species I offered an equal wet biomass of for 12 hr each day in addition to natural sunlight. each fruit speciesplaced randomly on the cage In captivity, the birds were fed domestic fruits floor at 08:OOhours every day. (apples, peaches, grapes, watermelons, and or- Each day I measured the total wet mass of the anges), vegetables (tomatoes, cucumbers, and fresh fruit provided, the total wet mass of fruits peppers), eggs,bread, and invertebrates (meal- left uneaten from the previous day’s provision, worms and locusts) for several days, until their the total wet mass of the excreted seeds,and the body mass stabilized, usually after lo-20 days. total wet massof the non-seed excreta (for a more There was no difference between the birds’ mass detailed description of this procedure see Izhaki when captured (37.7 ? 2.6 g) and on the first day and Safriel 1989). The non-seed excreta were of experiment (36.6 * 2.8 g, paired t-test t2, = oven-dried until they reached a constant mass 1.06, P > 0.05). and reweighed to obtain the dry mass of excreta Fecal analysis of 14 bulbuls captured in Beit per day (Qe). I weighed the birds daily at 08:OO Jimal, Judean Hills, Israel, between July and hr to determine whether or not to continue the February demonstrated that the birds ate five experiment. Trials lasted five to 10 days each. fruit species(Rubia tenuifolia, Asparagusaphyl- At the end of each trial, the bird was fed on the lus, Smilax aspera, Rhamnus palaestinus, and pre-experiment mixed diet. Its mass was moni- Pistacia lentiscus)with an average of 1.4 * 0.9 tored at five-day intervals for an additional 30 fruit speciesper bird (Izhaki 1986). Fecal analysis days. of 49 bulbuls captured at Ramat Hanadiv, Israel, between November and January revealed that NUTRITIONAL ANALYSIS the birds consumed five fruit species(Phillyrea Many fresh fruits (usually hundreds) were col- latifolia, A. aphyllus, Ephedra aphylla, R. pa- lected from each species.Fresh pulp from laestinus,P. lentiscus)with an average of 1.6 * the fruits were mixed, dried and divided into two 0.7 fruit speciesper bird (Adar 199 1). Two fruit samples. These samples of dried pulp from the speciesidentified by this analysis and one con- four fruit species and the dry feces from each generic specieswere used in the present study. experiment were analyzed for lipids, nitrogen, Becausebulbuls are frequently observed eating reducing ,ash, and several cations (Na, K, Lonicera fruits around our campus and in other Ca, P), according to AOAC (1975) procedures. locationsin Israel (Izhaki, pers. observ.), this fruit I estimated protein content of all speciesexcept was added to the diet. The data on fecal analyses Ephedra as total N x 6.25 (Bondi 1987). Ephed- covers a period of several months, but bulbuls ra is well known for its relatively high content actually consume fewer fruit speciesover shorter of alkaloids such as ephedrines (Herrera, pers. periodsof time, dependingon availability. Hence, comm.). Such nonprotein nitrogenous com- the four fruit speciesfed to bulbuls in this study pounds may cause overestimation of protein representgreater fruit diversity than exists in na- content when the traditional 6.25 factor is used ture. The four fleshy fruit species(madder, Rubia and so the protein content of Ephedra fruits was tenuifolia; buckthorn, Rhamnus alaternus;joint analyzed separately using Biotronic LC 5000 pine Ephedra aphylla; and honeysuckle, Lonic- Amino Acid Analyzer (Moore and Stein 195 1). era etrusca) used in feeding trials were picked This indicated that the correct conversion factor 914 ID0 IZHAIU for Ephedra fruits was 4.3, which was close to RESULTS 4.4, suggestedby Milton and Dintzis (198 1) for NUTRITIONAL CONTENT OF FRUITS tropical plants. I also measured the sugarcontent in the fruit juice in the field using a hand refrac- The four fruit specieshad a juicy pulp K = 70.3% tometer. The caloric content of dry pulps and + 7.8% water content), which was relatively poor feceswas estimated using the average grossen- in crude protein (4.6% + 1.6% of dry mass) and ergy equivalents of protein (17.2 kJ/g), fat (38.9 rich in carbohydrates (> 72% of dry mass, Table W/g) and carbohydrates (17.2 kJ/g, see Karlson 1). However, the pulps of Rhamnus and Ephed- 1972). ra contained almost twice as much protein as the pulps of Rubia and Lonicera. The pulp of METABOLIZABILITY COEFFICIENTS Rubia was relatively rich in lipids but relatively I calculated the Apparent Metabolizability (AM) poor in carbohydrates, and had no sugarsin its for each dietary constituent x, AMx by AMx = juice (Table 1). This pulp also contained a high (QiX - QEX)/QiX(Robbins 1983) where Qi. and concentration of energy (20.1 kJ/g) compared to Qe, are the quantity of substance x in the dry the mean for the other three fruit speciestogether consumed feed (intake) and in the dry excreta (16.5 kJ/g). All four fruit specieshad relatively produced, respectively. Because fecal and uri- low protein to energy ratios. While the total ash nary wastes cannot be separatedin birds I used content was relatively similar among the species the term “Apparent Metabolizability” (Robbins it was notable that Rubia fruits were exceptional 1983, Miller and Reinecke 1984). The Apparent in their low K to Na ratio and their high Ca to Metabolizable Energy (AME) was calculated by P ratio (Table 1). AME = ([GEi][Qi] - [GEe][Qe])/(GEi)(Qi), where GEi and GEe are the grossenergy density of the SEED LOAD consumed dry food and excreta produced, re- Bulbuls did not regurgitate seedsof the experi- spectively (Kendeigh et al. 1977). mental fruit species, but passed them through PROTEIN AND ENERGY REQUIREMENTS their digestivetracts undigested. The load of seeds The expected nitrogen intake for maintenance relative to nutritional gain of pulp was expressed by the ratio of seedmass to fruit mass (Table 1). was calculated from the equation, Nreqmred= 0.43 g N kg-o.75day-l (Robbins 1981, 1983) and ni- It is notable that Ephedra and Lonicera had a trogen values were converted to protein by mul- relatively low seedload (12-l 3%) compared with tiplying the correct conversion factor, as de- the other two species. scribed. The expected energy requirements for maintenance were calculated from the equation, FRUIT PREFERENCE E requ,red= 1.572 kgo.62(Kendeigh 1970). Fruit preference(calculated from pulp consump- tion) among the four fruit speciesthat were of- STATISTICAL ANALYSIS fered simultaneously varied among the individ- Two-factor analysis of variance (ANOVA) was uals but lessamong trial days (Fig. 1). The average used to examine the effectsof the four different pulp consumption of Lonicera and Ephedra fruits diets (one mixed and three diets of a single fruit (39% ? 9%, 32% +- 5%, respectively) was two species)and the effectsof the trial day (six days) to three times greater than pulp consumption of on fruit intake, relative to body mass, and on the Rhamnus (13% f 4%) and Rubia (17% + 6%) number of fruits eaten per day. One-way ANO- fruits. Four individuals (Nos. 3,4,5,6) preferred VA was used to detect the effect of the four diets Lonicera fruits while the other two almost ig- on mass loss, energy and protein consumption, nored them (Fig. 1). One of these two (No. 2) and metabolizabilities. For this analysis, I pooled preferred Ephedra fruits and the other (No. 1) the results for each individual over the first six preferred Rubia and Ephedra fruits (Fig. 1). Al- days of the trial. I used the REGWF Multiple F though the birds were relatively consistentin iheir test (SAS 1988) as an a posteriori test for com- main preference during the trial, an important parisons of these means. All proportions were part of their diet consistedof the other three fruit arcsine square-root transformed before analysis species(Fig. 1). It is notable that the birds pre- (Sokal and Rohlf 198 1). Values are presented as ferred the two fruit specieswith the lower seed means f standard deviations. mass to fruit mass ratio (Table 1). NUTRITIONAL QUALITY OF FRUIT DIET 915

TABLE 1. Morphological characteristics(Mean + SD) and nutritional content of the four fruit speciesfed to bulbuls.

Rubm tenuifolra Rhamnusalalernus Loniceraetruca Ephedraaphylia Morphological characteristics Sample sizea 147 26 52 25 No. of seedsper fruit 1.0 (0) 3.0 (0) 1.86 (1.09) 1.56 (0.5) Wet fruit mass (g) 0.16 (0.02) 0.26 (0.06) 0.15 (0.03) 0.20 (0.07) Wet pulp mass (g) 0.10 (0.01) 0.19 (0.06) 0.13 (0.02) 0.18 (0.07) Seed mass/fruit mass 0.41 (0.04) 0.28 (0.09) 0.13 (0.05) 0.12 (0.07) Nutritional content Water (O/o)” 59.0 72.0 73.0 77.0 Protein (O/o) 3.2 5.8 3.3 6.1 Fat (O/o)’ 18.1 0.1 4.1 0.7 Carbohydrates(O/o)” 72.8 88.5 87.9 83.7 Sugars(O/o) 0 28.2 20.9 23.1 K (%) 1.77 2.75 3.7 2.73 Na (o/o)’ 0.26 0.1 0.29 0.12 Ca (%) 0.92 0.34 0.58 0.29 P (%) 0.02 0.05 0.16 0.11 Ash (O/o)’ 5.9 5.0 4.73 6.7 WNa 6.8 27.5 12.8 22.8 Ca/P 46 6.8 3.6 2.6 Energy @J/g dry pulp)’ 20.1 16.4 17.3 15.7 =Number of fruits (a sampleof at least IO fruits from eachplant speciesused in feedingtrials). DPercentage of wet pulp mass. ePercentage of dry pulp mass. d Calculatedby subtractingthe percentagesof fat, protein, and ash from 100%. cg/100 g solution of fruit Juice. ‘Calculated by multiplying the amount of fat, protein, and carbohydratesm I gdry pulp by their energycontents (see Methods).

BODY MASS CHANGES the trial (Fig. 2). Thus, at the beginning of the Birds on all four diet regimens lost mass steadily experiments fruit mass consumption increased during the first seven days of the experiments each day as body mass decreased.The mean dai- (Fig. 2). Although the averagebody mass loss for ly rate of fruit intake (across individuals) was the entire trial acrossindividuals differed among significantly dependent on both the type of diet the four diets (one-way analysis of variance F3,z0 and the day of the trial (Table 2) with the highest = 4.72, P = 0.01) the average body mass loss of values for the Rubia diet (1.32 1?I0.20) and much birds fed on a mixed diet of fruits (- 10.8% ? lower values for the mixed fruit diet (0.68 -t 6.3%) was not significantly different from that of 0.09), the Rhamnus diet (0.68 f 0.06), and the birds fed on Rharnnus (- 7.2% f 3.8%), Rubia Ephedra diet (0.56 +- 0.11). The average daily (-14.2%+3.1%),andEphedra(-15.3%+2.2%, number of fruits consumed was 87 + 20 for REGWF Multiple F Test, P > 0.05). Rhamnus, 250 + 99 for Rubia, 102 k 24 for The daily proportion of body mass to pre-ex- Ephedra and 116 rf: 24 for the mixed diet. The periment body mass depended on the number of number of fruits consumed per day was signifi- days elapsed from the beginning of the trial, as cantly affected by both the type of diet and the indicated by the regression equations (Fig. 2). day of trial (Table 2). There were no significant differencesbetween any pair of regressioncoefficients in Fig. 2 (t > 0.1, PROTEIN AND ENERGY CONSUMPTION P > 0.1; Marascuilo and Serlin 1988:623) in- The average protein content of fruit pulps for dicating that there was no difference among diets bulbuls fed on mixed diet (4.6% f 1.6%) was in the rates of mass loss during the trial. lower than in exclusive diets of Rhamnus and Ephedra but higher than in diets of Rubia and FRUIT INTAKE RATE Lonicera (Table 1). Considering the amount of In all four diets the birds gradually increased food eaten per day, bulbuls fed on a mixed diet their fruit intake rate (wet mass of consumed consumed less protein (mg/g body mass per 6 fruits per day/body mass) until the third day of days) than those on the single fruit diets of 916 ID0 IZHAIU

1.0

0.8

.3S 0.6 [-c”l Lmiceru t: & e 0.4 Rubia a Ephedru 0.2

Day of Trial

I Zmicera m Rhamnw

Rubia Ephedru

1 2 3 4 5 6 Bird Number FIGURE 1. Fruit consumption of bulbuls fed a mixed diet of four fruit species during a seven day trial. Proportions are of total pulp’s wet mass consumption. The upper figure presentsthe mean proportions per day and the lower presentsthe mean proportions per individual over the seven day trial.

TABLE 2. Two-way ANOVA tables for the effect of Ephedru and Rubia (Table 3). In all trials, birds diet and day of trial on the number of fruits eaten per consumed more protein than their expected day, and on fruit intake (total wet fruit massconsumed maintenance requirement according to Robbins per day/body mass). Original data were transformed. (198 1, 1983) (Table 3). Bulbuls fed on a mixed diet assimilated less protein than those fed on Sourceof Meall variation df SLJ”X.5 F P Rubia and Rhamnus but more than those fed on Fruit intake Ephedru (Table 3). Bulbuls fed on Ephedra had Diet : 4.66 166.67 0.0001 a negative protein balance (Table 3). Day 0.13 4.72 0.0002 The average energy content of the mixed pulps Diet x day 18 0.12 4.72 0.0001 in bulbuls fed on a mixed diet (17.38 k 1.93 kJ/ Residual 129 0.03 g dry mass) was lower in birds fed on Rubia, Number of fruits eaten higher than in birds fed on Rhamnus and Ephed- Diet 3 224,457.6 85.46 0.001 ru and similar to that in bulbuls fed exclusively Day 6 7,201.4 2.74 0.015 on Lonicera fruits (Table 1). Accounting for the Diet x day 18 1,975.2 0.75 0.75 Residual 135 2,626.5 amount of pulp eaten per day, bulbuls on a mixed diet consumed less energy (kJ/g body mass per NUTRITIONAL QUALITY OF FRUIT DIET 917

2.0 1.8 MIXED DIET 1.6 1.4 1.2 I.21 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 1.00

0.94

0.88

0.82 t I

z 4

:: 3 2.0 2.0, I c3 T- 1.8 T 1.8 EPHEDRA 1.6 2 1.6 1 ti 1.4 1.4 Lr 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 1.00 0.2 1.00

0.94 0.94

0.88 0.88

0.82 0.82

0.76

0.70 1234567

DAY FIGURE 2. The rate of body mass loss (bottom) and fruit intake rate (top, pulp consumption/body mass) of bulbuls during the first seven days of feeding trial of a mixed fruit diet and three diets of a single fruit species. The regressionequations of body mass loss are also given. day) than birds fed on Rubia, more energy than The average dry matter metabolizability in birds eating only Ephedra, and same amount of birds consuming a mixed diet was significantly energy as birds fed on Rhamnus fruits (Table 3). higher than in those fed on Ephedra, lower than However, in all diets birds consumed more en- those fed on Rhamnus but with no difference ergy than expected based on their daily require- from those fed on Rubia (Table 4). The AM of ments (Kendeigh 1970). Bulbuls fed on Rubia dry matter of birds fed on a mixed diet decreased consumed 3.6 times more energy than their daily during the trial (average AM for the first three requirements and assimilated significantly more days of the trial was 0.76 and for the next three energy than those fed on the other three diets dayswas 0.66, paired t-test, t, = 7.72, P -C0.001). (Table 3). Bulbuls fed on Ephedra consumed only This pattern was not detected when birds were 8% more energy than required (Table 3). fed on exclusive diets of one fruit species(paired t-test for the three trials, P > 0.1). APPARENT METABOLIZABILITY (AM) Apparent metabolizabilities for birds eating the Bulbuls fed a mixed diet of fruits had a signifi- mixed fruit diet were in the range of values for cantly higher AME than those fed an exclusive birds fed on the exclusive diet of one fruit species diet of Rubia and Ephedra but lower than those (Table 4). Birds fed a mixed fruit diet had a lower fed on Rhamnus (Table 4). AM for protein than those fed an exclusive diet 918 ID0 IZHAKI

TABLE 3. The requirements of protein and energy for the maintenance of bulbuls vs. consumption when fed a mixed diet and singlefruit diets. One-way ANOVA testsfor the differencesamong different diets were executed on arcsine square-root transformed proportions. The results of REGWF multiple F test (P < 0.05, SAS 1988) among diets are indicated in each row by the superscriptsa-d: a > b > c 1 d. n = 6 bulbuls for each diet. The results are pooled from the first six days of the trials. Standard deviations are in parentheses.

Diet Mixed Rhamnus Rubm Ephedra F

Protein consumption (mg per 6 days/g body mass) 39.27 (3.96)” 46.31 (8.22)” 59.65 (6.12) 42.40 (4.87)” 13.36* Protein consumption/ expected protein intake+ 1.04 (0.10) 1.30 (O.lS)b 1.57 (0.18) 1.67 (0.18) 19.90* Protein assimilation (mg per 6 days/g body mass) 3.60 (5.89)b 24.27 (7.36)’ 24.84 (5.84) -45.61 (7.52) 146.37* Energy consumption (kJ per 6 days/g body mass) 16.10 (2.59)b 13.09 (2.33)bc 37.47 (3.84) 10.91 (1.25) 126.19* Energy consumption/ expected energy intake11 1.54 (0.24)” 1.33 (0.16)bC 3.55 (0.43)” 1.08 (0.11) 107.61* Energy assimilation (kJ per 6 days/g body mass) 11.66 (1.62)b 10.70 (2.15) 19.09 (2.72) 6.66 (0.89)’ 41.72*

* = P < 0.001. $ Expectednitrogen intake wascalculated by 0.43 g N kg- o.i day I (Robbins 1983) convertedto protein by multiplying by 6.25 (Bondi 1987) for mixed diet, Rhamnus and Rubm and by 4.3 for Ephedra (see Methods). 1)Expected energy intake was calculatedaccording to Kendeigh(1970). ofRhamnus and Rubia but higher than those fed low AMs were found for various components in on Ephedra (Table 4). Birds fed on Ephedra ex- different diets (Table 4). perienced a negative balance of proteins, , Na, Ca, and P (Table 4). Although negative bal- D*SCUSS1oN antes were not observed for the other diets (ex- The results of recent studies of the nutritional cept for phosphorusin the mixed diet), relatively value of fruits to frugivorous birds indicate that

TABLE 4. Apparent metabolizable energy(AME) and apparentdigestibilities of severalpulp constituents(AM) consumed by bulbuls. One-way ANOVA tests for the differencesbetween birds fed mixed diets and those fed a singlefruit diet were executedon arcsinesquare-root transformed proportions.The resultsof REGWF Multiple F test (P < 0.05) among diets are indicated in each row by the superscriptsa-d: a > b > c > d. n = 6 bulbuls for each diet. The resultsare pooled from the first six days of the trials. Standard deviations are in parentheses.

Diet Mixed Rhamnus Rubia Ephedra F

AME$ 0.73b (0.03) 0.82 (0.02) 0.69 (0.02) 0.61d (0.03) 73.44* AMII Dry matter 0.71b (0.02) 0.83a(0.03) 0.72b (0.01) 0.63c (0.03) 82.48* Carbohydrates 0.78b (0.01) 0.85’ (0.03) 0.78b (0.01) 0.78b (0.02) 16.99* Proteins 0.03b (0.22) 0.55a (0.11) 0.41a (0.07) -1.21c(0.15) 176.12* Fats 0.6 la (0.08) 0.26a (0.10) 0.53a(0.04) - 1.25” (0.79) 28.26* Ash 0.34” (0.06) 0.59a (0.09) 0.59 (0.04) 0.02 (0.08) 82.77* Na 0.61a (0.18) 0.35b (0.18) 0.74 (0.05) -l.l@ (0.21) 167.82* K 0.52b (0.18) 0.97’ (0.01) 0.55b (0.02) 0.07c (0.07) 87.55* Ca 0.31b(0.13) 0.62a(0.08) 0.66a (0.03) -0.27c (0.12) 112.32* P _-0.55c (0.75) 0.22b (0.11) 0.77a(0.08) -1.73d(0.18) 46.25*

* = P < 0.001. t (Energyconsumed - energyexcreted)/energy consumed. 1:(Constituent x consumed constituentx excreted)/constituentx consumed NUTRITIONAL QUALITY OF FRUIT DIET 919

most birds are unable to subsist on an exclusive and the requisite protein content of fruits to sat- diet of fruits (Berthold 1976; Bairlein 1987; Her- isfy these demands, are not yet clear. Robbins rera 1984; Izhaki and Safriel 1989, 1990a; Levey (198 1, 1983) provides a general equation for ni- and Karasov 1989; but see Walsberg 1975; Hol- trogen intake required for nitrogen balance. thuijzen and Adkinsson 1984; Jordan0 1988; However, several frugivorous birds, including Studier et al. 1988; Worthington 1989). In this bulbuls, consume more protein than expected and another study (Levey and Karasov 1989) but assimilate lessprotein than required accord- where a mixed fruit diet was presented,the birds ing to Robbins’ equation (Izhaki and Safiiel 1989, lost mass (but see Holthuijzen and Adkisson Levey and Karasov 1989). Relatively low protein 1984). The rate of body mass loss (Fig. 2) for digestion was noted in this study as well (Table birds fed on a mixed fruit diet did not differ from 4). Thus, the low protein content of the fruits that for birds fed on a single fruit diet. was not compensated for by an increase in effi- ciency of protein digestion. However, a mixed THE PROTEIN DEFICIENCY OF FRUITS fruit diet improved the birds’ ability to digest The average low protein content (4.6%) of the and assimilate more protein as compared with four fruit speciesused in this study doesnot differ an exclusive diet of Ephedra but not exclusive substantially from those of other regions such as diets of Rubia and Rhamnus (Table 3). the Iberian Peninsula (5.1% * 3.8%, n = 92 [Her- It is notable that the average protein content rera 19871) southern England (6.1% * 5.4%, y1 of fruits required for nitrogen balance is not sig- = 28 [Snow and Snow 19881, central Sweden nificantly different from the averageprotein con- (5.2% ? 3.8%, n = 3 1 [Eriksson and Ehrlen tent leading to protein imbalance (Table 5, t,, = 19911) central Illinois (5.1% ? 2.3%, y1= 20 0.77, P > 0.5). However, because the data in- [Johnsonet al. 1985]), Washington (5.3% k 3.4%, clude results of experiments undertaken in dif- n = 18 [Piper 19861) and Hawaii (7.0% t 3.5%, ferent conditions and from birds of different n = 22 [Sakai and Carpenter 19901). Further- regions (temperate and tropical), this conclusion more, it is likely that the actual available protein must be viewed with caution. Nevertheless, the for birds is even lessthan thesevalues since pulps data indicate that the protein content alone tells may contain non-digestible nitrogen, such as in us little about the ability of birds to subsist on secondarycompounds (Milton and Dintzis 198 1; fruits. Ephedra fruits with a relatively high pro- Sedinger 1990; and see Methods). tein content (Table 1) were the only ones which Protein deficiency has been suggestedas a fac- causeda negative protein balance (Table 3). Bair- tor explaining the body mass loss and nitrogen lein (1987) reported that the frugivorous Sylvia imbalance in cage experiments (Moss and Par- borin fed on an artificial diet that was very low kinson 1975, Berthold 1976, Bairlein 1987, Le- in protein (2.4% wet mass) lost weight imme- vey and Karasov 1989, Sedinger 1990). Birds diately at the beginning of the trial, but later may cope with this protein shortage either by maintained a constant mass. Costa’s Humming- switchingto insectivory (e.g., Morton 1973, Bert- birds (Calypte costae) can maintain their body hold 1976, Foster 1978,LeveyandKarasov 1989) mass on only 1.5% protein in their diet (Brice or by increasing their rate of fruit intake (but see and Grau 199 1). Low protein diets in these cases Johnson et al. 1985, Bairlein 1987) and gut pas- may be tolerable if birds can consume and digest sage rate as long as high digestion efficiency is a large volume of food. However, these obser- maintained. This second solution probably vations may indicate that a factor other than evolved in manakins (Worthington 1989) but was protein concentration per se may regulate body undetected in several temperate species where mass loss in bulbuls. short retention times actually decreasefruit di- Fruit pulp may lack some specificamino acids gestion (Izhaki and Safriel 1989, Levey and Ka- (Parrish and Martin 1977, Mack 1990, Sedinger rasov 1989). The main problem might not be the 1990) and birds may actually select their diets protein shortage per se, but the relatively high on the basis of their amino acid contents (e.g., ratio of energy to protein in pulps. This results Murphy and King 1989). In arils of Aglaia spe- in the fulfillment of energy demands before pro- cies that were analyzed, only two species con- tein demands (Ricklefs 1976, Foster 1978, Sibly tained arginine, two species contained cystine 198 1, Jordan0 1988). However, the protein re- and none contained histidine (Pannell and Ko- quirements for maintenance of frugivorous birds, ziol 1987). There are no data on the amino acid 920 ID0 IZHAKI

TABLE 5. The ability of frugivorousbird speciesto maintain a positive nitrogen balance as a function of pulp protein content.

Protein content of fruit pulp (% dry mass)t Positwe Negative Bird species Fruit species N balance N balance Reference

Phainopeplanitens Phorandedroncalifornicum 1.5 Walsberg 1975 Sturnusvulgaris* Cornusracemosa 6.1 Levey and Karasov 1989 Viburnum dentatum Vitis sp. Turdus migratorius$ Cornusracemosa 6.1 Levey and Karasov 1989 Viburnum dentatum Vitis sp. Manacus vitellinus Heliconia latispatha 8.5 Worthington 1989 Psychotriadejlexa 8.1 Hasseltiajloribunda 7.1 Psychotriahorizontalis 6.3 Byrsonimacrassifolia 5.1 Anthurium brownii 2.6 Pipra mentalis Anthurium clavigerum 8.3 Worthington 1989 HasseltriaJloribunda 7.1 Psychotriamarginata 6.3 Psychotriahorizontalis 6.3 Pycnonotusxanthopygos Rhamnus alaternus 3.3 Present study Rubia tenuifolia 3.2 Ephedra aphylla 6.1 A mixed diet of the three speciesabove and Lonicera etruscal: 4.6 Bombycillacedrorum 11 Crataegusphaenopyrum 9.2 Studier et al. 1988 Mean +- standard deviation 6.6 k 2.3 5.9 + 1.4

t Nitrogen converted to protein by multiplying by 6.25 except for Ephedra, where the 4.3 factor was used (see Methods) $ A diet of mixed fruit species, the value of protein content is the average for these fruit species. I(Data for free living birds.

contents of the pulps in the present study, but if of secondary compounds. After consuming this bulbuls suffered from a shortageof amino acids amount they move to another species(Freeland when fed a single species of fruit they did not and Janzen 1974). It is possible however, that improve amino acid balance when fed a mixed frugivorous birds have evolved mechanisms of diet. detoxification (Herrera 1985). For example, the Several fruit speciescongeneric with thoseused relatively large livers found in starlings might as food in my trials (Rhamnus lycioides,Rubia help them to detoxify secondary compounds peregrina, Lonicera periclymenum) contain sec- (Moermond and Denslow 1985). ondary compounds (such as glucosids, an- thrones, saponins) in their ripe pulp (Jordan0 THE AME PROBLEM 1988). Secondary compounds in pulps may re- An interesting finding in this and similar studies duce the availability of protein in the diet and (Holthuijzen and Adkinsson 1984, Izhaki and the activity of digestive enzymes (Feeny 1969; Safriel 1989, Levey and Karasov 1989) are the Swain 1979; Robbins 1983; Robbins et al. 1987; low AME values. If frugivorous birds accelerate Bemays et al. 1989; Izhaki and Safriel 1989, food intake rate to maximize protein intake, they 1990a; Levey and Karasov 1989; but see Mack may also reduce the efficiency of energy assim- 1990; Sedinger 1990). A bird possibly can shift ilation (Izhaki and Safiiell989, Levey and Kara- from one fruit to another to avoid ingesting too sov 1989). Levey and Karasov (1989) however, high levels of a single secondarycompound (Jor- believed that low AME values were responsible dano 1988, Levey and Karasov 1989, Mack for nitrogen imbalance through catabolism of 1990). This is similar to the behavior suggested their body protein to meet energy requirements. for herbivores.They may tolerate a small amount Izhaki and Safiiel(l989,1990a), in contrast, sug- NUTRITIONAL QUALITY OF FRUIT DIET 921 gested that low protein assimilation caused in potassium to sodium ratio as high as 20 (Bondi part by secondary compounds leads to acceler- 1987). It is still unknown if and how this high ation of the gut passagetime and hence to low potassium to sodium ratio is related to the poor AME values. Koenig (199 1) found that adding physiological state of bulbuls fed on an exclusive tannins to acorn meal eaten by Acorn Wood- diet of fruits with unlimited water. peckers (Melunerpes formicivorus) decreased CONCLUSIONS AME. It should be emphasizedthat the low AME This study does not confirm the hypothesis that values reported by Levey and Karasov (1989) a mixed fruit diet improves bird maintenance in were also found when birds were fed on diets captivity better than a single fruit diet. Birds ap- without secondary compounds. However, the parently have tried to compensate for the low main advantage of a mixed fruit over a single protein levels of the fruits primarily by increasing fruit diet of Rubia and Ephedra for bulbuls was daily fruit intake. However, becausethey did not their relatively high AME (Table 4). These AME increase the efficiency of protein digestion they values, though, are still below those expectedfor lost weight at the same rate, regardlessof their fiugivorous birds (seeLevey and Karasov 1989). diet. Birds selectedthe fruits with the lowest seeds MINERAL DEFICIENCY OF FRUITS to pulp load (Lonicera and Ephedra) and not on the basisof any nutritional attribute. Bulbuls may Frugivorous birds may lose mass because of a have consumed large amounts of fruit in order severeshortage of minerals in pulps (e.g., sodium to cope with other pulp attributes such as high and potassium). These function with phosphates energy to protein ratio and high K to Na ratio. and bicarbonates to maintain homeostasis, os- I hypothesizethat such a high fruit consumption motic relationships and optimum pH of the body and relatively low digestibility may lead to (1) (Simons 1986, Bondi 1987). Sodium lack may decreasing AME, (2) imbalance in mineral nu- cause an immediate mass loss (Moss and Par- trition and (3) accumulation of secondary com- kinson 197 5). A negative balance of sodium was pounds to levels that interfere with protein di- (Bom- observed in free-living Cedar Waxwings gestion. These factors may lead to body mass bycilla cedrorum) feeding on Washington Haw- loss thorn fruits (Studier et al. 1988) but not in the This ’ study does not provide direct evidence bulbuls in this study (Table 4). for the “secondary compounds hypothesis” and There are no data on the mineral requirements further study on the actual content of secondary of frugivorous birds. The only available data for compounds in these fruits is needed. An alter- birds is for domestic forms. Pheasants (Phasi- native explanation for the bouts of insectivory anus colchicus)and JapaneseQuail (Coturnix co- in frugivorous birds is the shortageof the same turnix japonica) need 0.15% sodium in their free amino acids in all ripe fruit species(Mack foodstuff for growth and breeding (National Re- 1990, Sedinger 1990). Following this line one search Council 1984). The requirements of all needs to analyze the amino acid content of fruits domestic animals for sodium are of the order of and to compare it with the birds’ requirements 0.1%-0.2% of dry matter (Bondi 1987). Thus, if for maintenance the sodium demands of the bulbuls are of the same order, the sodium content of the four fruit ACKNoWLEDGMENTS species (0.19% ? O.l%, Table 1) should fulfill I gratefully acknowledgethe critical comments of C. their demands. The four fruit specieshave a high M. Herrera, P. Jordano, D. J. Levey, A. L. Mack, U. ratio of potassium to sodium (18 f 9, Table 1). N. Safriel and J. S. Sedinger,which greatly improved this paper. I would also like to acknowledgeassistance An extreme predominance of potassium is de- from 0. Karnieli, H. Kuzik, I. Neeman, and T. Avivi. tected in fruits from other regions, such as the Nutritional analyseswere performed in Miloda Lab- Iberian Peninsula (84 ? 53, n = 72 [Herrera oratoriesunder D. Izhar’s supervision.The study was 1987]), Illinois (42 + 25, n = 20 [Johnson et al. partly supported by the Carmel Foundation, Israel Ministry of Environmental Affairs, and by an Oranim 1985]), and Hawaii (10.5 + 8.1, n = 22 [Sakai researchgrant. and Carpenter 19901). Japanese Quail needs a potassium to sodium ratio of 2.7 for growing and LITERATURE CITED breeding (National Research Council 1984). ADAR, M. 1991. The interactionsbetween Asparagus However, animals are able to regulate the level aphyllusand its dispersalagents. M.Sc.thesis, The of these two ions, and some can cope with a Hebrew Univ. of Jerusalem,Israel (in Hebrew). 922 ID0 IZHAKI

AL-DABBAGH,K. Y., J. H. JIAD, AND I. N. WAHEED. IZHAKI, I., AND U. N. SAFRIEL. 1990b. The effect of 1987. The influence ofdiet on the intestine length some Mediterranean scrubland frugivores upon of the White-cheeked Bulbul. Ornis Stand. 18: germination pattern. J. Ecol. 78:56-65. 150-152. IZHAKI, I., AND U. N. SAFRIEL. 1991. Seed shadow AOAC. 1975. Official methods of analysisof the As- generatedby frugivorousbirds in an easternMed- sociation of Official Analytical Chemists. AOAC, iterranean scrub. J. Ecol. 79:575-590. Washington, DC. JOHNSON,R. A., M. F. WILLSON,J. N. THOMPSON,AND BAIRLEIN, F. 1987. Nutritional requirements for R. I. BERTIN. 1985. Nutritional values of wild maintenanceof body weight and fat deposition in fruits and consumption by migrant frugivorous the long-distancemigratory Garden Warbler, Syl- birds. Ecology 66:8 19-827. via borin (Boddaert). Comp. Biochem. Physiol. JORDANO,P. 1988. Diet, fruit choice and variation 86A:337-347. in body condition of frugivorouswarblers in Med- BERNAYS,E. A., G. COOPERDRIEVER, AND M. BIL- iterranean scrubland.Ardea 76:193-209. GENER. 1989. Herbivores and plant tannins. Ad- KARLSON,P. 1972. Biochemie, 8th ed. Thieme, Stutt- vances in Ecol. Res. 19:263-30 1. gart, Germany. BERTHOLD,P. B. 1976. The control and significance KSNDEIGH,S. C. 1970. Energy requirements for ex- of animal and vegetable nutrition in omnivorous istence in relation to size of bird. Condor 72:60- songbirds.Ardea 64: 140-l 54. 65. BONDI, A. A. 1987. Animal nutrition. John Wiley KENDEIGH,S. C., V. R. DOLNIK,AND V. M. GAVRILOV. and Sons, New York, NY. 1971. Avian energetics, p. 127-204. In J. Pi- BRICE.A. T.. AND C. R. GRAU. 199 1. Protein re- nowski. and S. C. Kendeiah. Ieds.1.Granivorous quirement of Costa’s Hummingbirds Culypte cos- birds in ecosystems. Cambridge- Univ. Press, tae. Physiol. Zool. 64161 l-626. Cambridge. ERIKSSON,O., AND J. EHRLEN. 1991. Phenological KOENIG, D. K. 1991. The effect of tannins and lipids variation in fruit characteristicsin vertebrate-dis- on digestion of acorns by Acorn Woodpeckers. persed plants. Oecologia 86:463-470. Auk 108:79-88. FEENY,P. P. 1969. Inhibitory effectofoak leaftannins LAMBERT,F. 1989. Fig-eating birds in a Malaysian on the hydrolysis of proteins by trvpsin. Phyto- lowland rain forest. J. Trop. Ecol. 5:40 l-412. chemistry 8:1226-1229. _ __ _ LEVEY,D. J. 1986. Methods of seed processingby FOSTER,M. S. 1978. Total frugivory in tropical pas- birds and seeddeposition patterns,p. 147-158. In serines:a reappraisal.Trop. Ecol. 19:13 l-l 54. A. Estrada and T. H. Fleming [eds.], Frugivores FREELAND,W. J., AND D. H. JANZEN. 1974. Strategies and seed dispersal. W. Junk Publ., Netherlands. in herbivory by mammals: the role of plant sec- LEVEY.D. J.. AND W. H. KARASOV. 1989. Digestive ondary compounds. Am. Nat. 108:269-289. responsesof temperate birds switched to fruit or HERRERA,C. M. 1981. Are tropical fruits more re- insect diets. Auk 106:675-686. warding to dispersersthan temperate ones?Am. MACK, A. L. 1990. Is frugivory limited by secondary Nat. 118:896-907. compounds in fruits? Oikos 57: 135-l 38. HERRERA,C. M. 1984. Adaptation to frugivory of MARASCUILO,L. A., AND R. C. SERLIN. 1988. Statis- Mediterranean avian seed dispersers.Ecology 65: tical Methods for the Social and Behavioral Sci- 609-617. ences. W. H. Freeman, New York, NY. HERRERA,C. M. 1985. Aposematic insects as six- MEINERTZHAGEN,R. 1954. Birds of Arabia. Oliver leggedfruits: incidental short-circuiting their de- and Boys, Edinburgh and London. fense by frugivorous birds. Am. Nat. 126:286- MILLER, M. R., AND K. J. REINECKE. 1984. Proper 293. expression of metabolizable energy in avian en- HERRERA,C. M. 1987. Vertebrate-dispersed plants ergetics.Condor 86:396-400. of the Iberian Peninsula: a study of fruit charac- MILTON, K, AND R. D. DINTZIS. 1981. Nitrogen-to- teristics. Ecol. Monoer. 57:305-33 1. protein factors for tropical plant samples.Biotro- HOLTHULIZEN,M. A., AND?. S. ADKISSON.1984. Pas- pica 13:177-181. sagerate, energetics,and utilization efficiency of MOERMOND,T. C., AND J. S. DENSLOW. 1985. Neo- the Cedar Waxwing. Wilson Bull. 96:680-684. tropical frugivores: patterns of behavior, mor- IZHAKI, I. 1986. Seed dispersalby birds in an eastern phology and nutrition with consequencesfor fruit Mediterranean scrub. Ph.D.diss. The Hebrew selection, p. 865-897. In P. A. Buckley, M. S. Univ. of Jerusalem,Israel (in Hebrew). Foster, E. M. Morton, R. S. Ridgely, and N. G. IZHAKI,I., AND U. N. SAFRIEL. 1985. Why do fleshy- Buckley [eds.]. Neotropical Ornithology, Number fruit plants of the Mediterranean scrub intercept 36. American Ornithologists’ Union, Washington, fall-but not spring-passage of seed-dispersing DC. migratory birds? Oecologia 67:40-43. MOORE,S., AND W. H. STEIN. 1951. Chromatography IZHAKI-I., AAD U. N. SAF~E~. 1989. Why are there of amino acids on sulfonated polystyreneresins. so few exclusivelvfmaivorous birds?Experiments J. Biol. Chem. 192:633. on fruit digestibility. &kos 54:23-32. _ MOREAU,R. E. 1972. The palaearctic-African bird IZHAKI, I., AND U. N. SAFRIEL. 1990a. Weight losses migration system. Academic Press, London. due to exclusive fruit diet-interpretation and MORTON,E. S. 1973. On the evolutionary advantages evolutionary implications: a reply to Mack and and disadvantagesof fruit eating in tropical birds. Sedinger.Oikos 57: 140-142. Am. Nat. 107:8-22. NUTRITIONAL QUALITY OF FRUIT DIET 923

Moss, R., AND J. A. PARKINSON.1975. The digestion SEDINGER,J. S. 1990. Are plant secondary com- of bulbils (Polygonurn viviparum) and berries pounds responsible for negative apparent meta- (Vaccinium myrtillus L. and Empetrum sp.) by bolizability of fruits by passerinebirds? A com- captive ptarmigan (Lagopus mu&s). Br. J. Nutr. ment on Izhaki and Safriel. Oikos 57: 138-140. 33:197-206. SIBLY.R. M. 1981. Strateaiesof digestion and def- MURPHY, M. E., AND J. R. KING. 1989. Sparrows ecation. p. 109-139. Inc. R. Townsend and P. discriminate between diets differing in valine or Calow [eds.], Physiologicalecology: an evolution- lysine concentrations.Physiol. and Behav. 45:423- ary approachto resourceuse. SinauerAssoc., Sun- 430. derland, MA. NATIONALRESEARCH COUNCIL. 1984. Nutrient re- SIMONS,P. C. M. 1986. Major minerals in the nu- quirements of poultry. 8th ed. National Science trition of poultry, p. 141-l 54. In C. Fisher and K. Press, Washington, DC. N. Boorman [eds.],Nutrient requirementsofpoul- PANNELL,C. M., AND M. J. KOZIOL. 1987. Ecological try and nutritional research.Butterworths, Kent, and phytochemical diversity of arillate seeds in England. Galaia (Meliaceae):a study of vertebrate dispersal SNOW,B., AND D. SNOW. 1988. Birds and berries. in tropical trees. Phil. Trans. R. Sot. Lond. 36: Poyser, Calton, United Kingdom. 303-333. SOKAL,R. R., AND F. J. ROHLF. 1981. Biometry. W. PARRISH,J. W., AND E. W. MARTIN. 1977. The effect H. Freeman, San Francisco. of dietary lysine on the energy and nitrogen bal- STUDIER,E. H., E. J. SZUCH,T. M. THOMPKINS,AND ance of the Dark-eyed Junco. Condor 79~24-30. V. W. COPE. 1988. Nutritional budgets in free PIPER, J. K. 1986. Seasonalityof fruit charactersand flying birds: Cedar Waxwings (Bombycilla cedro- seed removal by birds. Oikos 46:303-3 10. rum) feeding on Washington hawthorn fruit (Cra- RICKLEFS,R. E. 1976. Growth rates of birds in the taegus phaenopyrum). Comp. Biochem. Physiol. humid New World trouics. Ibis 116:179-207. 89A47 l-474. ROBBINS,C. T. 198 1. EstLmationof the relative pro- SWAIN,T. 1979. Tannins and lignins, p. 657-682. In tein costof reproductionin birds. Condor 83: 177- G. A. Rosenthal and D. H. Janzen [eds.], Herbi- 179. vores: their interaction with secondaryplant me- ROBBINS,C. T. 1983. Wildlife feeding and nutrition. tabolites. Academic Press, New York, NY. Academic Press, New York, NY. WALSBERG.G. E. 1975. Digestive adautations of ROBBINS,C. T., T. A. HANLEY,A. E. HAGERMAN,0. Phainipepla nitens associatedwith the eating of HJELJORD,D. L. BAKER,C. C. SCHWARTY,AND W. mistletoe berries. Condor 77: 169-l 74. W. MAUTZ. 1987. Role of tannins in defending WHISTLER,H. 1949. Popular handbook of Indian plants against ruminants: reduction in protein birds. Gurney and Jackson, London and Edin- availability. Ecology 68:98-107. burgh. SAKAI,H. F., ANDJ. R. CARPENTER.1990. The variety WORTHINGTON,A. H. 1989. Adaptations for avian and nutritional value of food consumed by Ha- frugivory: assimilation efficiency and gut transit waiian Crow nestlings, an endangered species. time of Manacus vitellinus and Pipra mentalis. Condor 921220-228. Oecologia 80:38 l-389. SAS INSTITUTE,INC. 1988. SAS user’s guide. Cary, NC.