Effect of biologically active plants used as netst material and the derived benefit to starling nestlings

Transcription

1 USDA Natinal Wildlife Research Center Frm the SelectedWrks f Larry Clark 1988 Effect f bilgically active plants used as netst material and the derived benefit t starling nestlings Larry Clark J. R. Masn Available at:

2 Oeclgia (1988) 77: Oeclgia 9 Springer-Verlag 1988 Effect f bilgically active plants used as nest material and the derived benefit t starling nestlings Larry Clark 1 and J. Russell Masn 1' 2 1 Mnell Chemical Senses Center, 35 Market Street, Philadelphia, PA 1914, USA 2 USDA/APHIS/ADC, Denver Wildlife Research Center, Sectin f Bird Cntrl, Bldg. 16, Denver Federal Center, Denver, CO , USA Summary. The Eurpean starling Sturnus vulgaris preferentially incrprates fresh sprigs f particular plant species fr use as nesting material. Chemicals fund in these plants may act t reduce pathgen and ectparasite ppulatins nrmally fund in nest envirnments. The present experiments were perfrmed t test this Nest Prtectin Hypthesis. In the field, we experimentally determined that wild carrt Daucus carta, a plant species preferred as nest material, effectively reduced the number f hematphagus mites fund within nests relative t cntrl nests withut green vegetatin. Chicks frm nests cntaining wild carrt had higher levels f bld hemglbin than chicks frm cntrl nests. Hwever, there were n differences in weight r feather develpment. In the labratry, we fund that wild carrt and fleabane, Erigern philadelphicus, (als preferred by starlings as nest material) substantially reduced the emergence f feeding instars f mites, while garlic mustard, Alliaria fficinalis, (cmmnly available but nt preferred) had little effect n the emergence f mites. We infer that preferred plant material may act t inhibit feeding r therwise delay reprductin f mites, thereby reducing risk f anemia t develping nestlings. Key wrds: Sturnus vulgaris - Nesting behavir- Ectparasite - Nest prtectin hypthesis - Bilgical cntrl Increasingly, evidence indicates that heavy infestatin f nests by ectparasites can affect survivrship and fecundity f breeding adult birds. Fr example, seabirds and swallws abandn breeding clnies fllwing a build-up in ectparasite ppulatins (Hgland and Sherman 1976; Duffy 1983). In thse cases where breeding persists, high ectparasite numbers may result in depressed grwth f the yung, and by implicatin, a decrease in their lng-term, pstfledging survival prbabilities (Mss and Camin 1971; Arendt 1985 a, b; Brwn and Brwn 1986). Ectparasites als can cause physilgical stress, thus increasing the risk f pprtunistic infectin by pathgens, r serve as a direct vectr fr pernicius frms f pathgens (Hfstad et al. 1984). Birds have several ptins t minimize ectparasite lad. One ptin is the use f nest-sites fr nly ne breeding seasn. This strategy reduces the time during which ectparasites can clnize the nest. Birds that re-use nests Offprint requests t." L. Clark r nest-sites incur higher parasite lads relative t species that utilize nests r sites nly nce (Stner 1936; Rthschild and Clay 1957; Wasylik 1971; Lye 1985; Barclay 1988). Placement f a nest away frm ther nests als can reduce the clnizatin rate by ectparasites (Brwn and Brwn 1986). Neither f the abve tw ptins are available t sme species (e.g. secndary cavity nesters) fr whm nest-sites may be a limiting resurce (Hgstad ]975; Brush 1983; Brawn and Balda 1988). In this circumstance, even use f a different site every year des nt preclude previus use f that site; thus, escape frm parasites may nt be pssible. Increased nest sanitatin and preening may be helpful in reducing ectparasite lad, but additinal strategies may be necessary. Such strategies are suggested by the nest prtectin hypthesis. Sme species f birds incrprate fresh green vegetatin int their nest material. Nrmally, the quantities f material are small relative t the remainder f the structurally supprtive dry nest matrix. The naturally ccurring defensive chemicals cntained in the plants are hypthesized t be txins (either vlatile r cntact) which can reduce parasite and pathgen lad. Supprt fr the nest prtectin hypthesis has been anecdtal and speculative (Jhnstn and Hardy 1962; Sengupta 1981). Hwever, Wimberger (1984) and Clark and Masn (1985) shwed that falcnifrms and passerines that reused nest-sites acrss breeding seasns, and by implicatin, incurred high ectparasite lads, were mre likely t use green plant material in nest cnstructin. In the case f the Eurpean starling Sturnus vulgaris, the plants used were selected in quantities greater than their prprtinal availability frm the habitat. These same plants were mre likely t have antibacterial and insecticidal prperties than a randm subset f available vegetatin. The bjective f the present study was t prvide mre detailed infrmatin n the efficacy f sme f the plants used by starlings at reducing ectparasite infestatin f nests. Specifically, we determined hw effective wild carrt Daucus carta and fleabane Erigern philadelphicus, tw plants preferentially utilized by starlings during nest cnstructin, were at limiting the hematphagus mite Ornithnysus sylviarurn. Bth labratry and field studies were carried ut. A secnd bjective was t assess hw the treatment f nests with fresh plant material (and by implicatin, the manipulatin f parasite lad) affected the well-being f nestlings. We hpe t understand eventually the prcesses by which starlings acquire this cmpnent f nest

3 175 building behavir and t elucidate the cues birds use t chse amng plants. Methds Stu@ site The nesting clny f starlings was lcated at the Strud Water Research Center (SWRC) f the Academy f Natural Sciences f Philadelphia, Avndale, PA. The habitat surrunding SWRC is a mix f lawns, ld fields, pastures, and secndary grwth deciduus frests. Individual nest bxes used in this study were cnstructed in 198 f unifrm materials and t unifrm dimensins. Such unifrmity f nest bx cnstructin may be imprtant because there is reasn t suspect that larger quantities f green plants are used by starlings nesting in lder bxes (Tenvu and Lemmetyinen 197; L. Clark, unpublished). Bilgy f the nrthern fwl mite We chse t mnitr numbers f the nrthern fwl mite because this was the mst cnspicuus ectparasite ccurring at the nest clny ver the years. O. sylviarum is an ectparasite which feeds primarily n the bld f birds. O. sylviarum can cmplete a five stage life cycle in 7-14 days, while adult frms are capable f enduring fasts f up t 3-4 mnths (persnal bservatin). Sykes and Chamberlain (1954) reprt that eggs are laid n birds r in nest material. After 24 h, nnfeeding, six-legged larva hatch. After an additinal 24 h, larva metamrphse int an eight-legged feeding prtnymph. All life-histry stages f O. sylviarum remain in the nest matrix. Feeding frms ascend t the hst, feed, then return t the nest matrix. Nrmally, prtnymphs require at least tw separate feeding buts (36 h) befre they are fully engrged (.25 mg bld) and change int a nnfeeding deutnymph stage. After 24 h, deutnymphs give rise t a cntinuusly feeding adult frm. Adult females may engrge with.41 mg f bld every 1-5 days befre laying an average clutch f 2.27 eggs. This life histry allws a small funding ppulatin t expand t cnsiderable numbers during the 37-4 days starlings are active at the nest (nest cnstructin thrugh fledging f the chicks). Nest manipulatins We manipulated nests t test the hypthesis that green plants influenced ppulatin levels f mites. T cntrl fr previus infestatin histry all bxes were scrapped clean f nest debris during January 1984 and We daily remved all fresh green plant material frm nest-bxes and the dry nest matrix thrughut the nest-building seasn in March and April as an effrt t standardize cnditins. Further mnitring f the nests' matrix was unnecessary, because nce the eggs are laid, additin f green plants t the nest n lnger ccurs (Clark and Masn 1985). In 1984 nests were randmly assigned t ne f three treatment levels : 1. Nest replacement (NR). The entire dry matrix f nest material was remved and replaced with an equivalent clean mass f dry grass material n -1, -5, 1, 7 and 13 days pst-hatching. Starling nests are easily replicated by the investigatr if care is taken t select similar nest material. This treatment served t reduce mite ppulatins thrughut the nestling's develpment. 2. Plant remval (PR). Once fresh plant material was remved these nests received n further treatment. These nests were presumably devid f putative miticidal r repellent effects f plants (sensu Ambasta 198) and wuld hypthetically a/lw mite ppulatins t grw uncnstrained. 3. Plant additin (PA). Fresh plant material (5 g) was interwven int the dry nest matrix at - 1, - 5, 1, 7 and 13 days pst-hatching. Starlings differentially select several herbaceus species fr inclusin int the nest relative t available vegetatin. Wild carrt was selected because it had the highest preference rating (utilizatin:availability) f green plants used by starlings breeding in sutheastern Pennsylvania (Clark and Masn 1985). During 1984, 3 nests frm each f the 3 treatments were cllected in May and June when the nestling ccupants were 1, 7, 13 and 19 days f age (N=72). Nest materials and debris were placed in a ne galln 'zip-lck' plastic bag and transprted t the labratry. Because feeding stages f O. sylviarum demnstrated psitive getaxis, we were able t cllect individuals by aspiratin at 24 h intervals. After 4-5 days, this methd yielded asympttically fewer mites. Pieces f nest material were then placed in a prcelain pan and washed in 7% ethyl alchl. The entire nest was washed in an effrt t extract as many f the remaining mites as pssible. Mites were cllected using a bulb pipet. We fund that nly the prtnymph, deutnymph and adult frms culd be reliably cllected. The eggs were t small t be easily seen and the larva were t small and translucent t be reliably seen, even under sterescpic magnificatin and illuminatin. After cllectin, mites were srted accrding t life histry stage. Srted ppulatins were placed in a centrifuge tube and spun at 25 rpm fr 1 rain n a clinical centrifuge. Numbers f mites were determined by cmparing vlumes f samples t calibrated samples whse numbers and vlumes were previusly determined. Differences in the lg number f mites fund in nests were analyzed using a fixed effects analysis f variance mdel with 3 factrs: nest manipulatin (3 levels), age f chicks (4 levels) and mnth hatching ccurred (2 levels). Nestling grwth The grwth f chicks frm a randmly selected subset f manipulated nests was mnitred. During hatching (day ) nests were adjusted t cntain 4 chicks. Only nests cntaining 4 yung thrughut 18 days f develpment were included in the cmparisn f age-specific grwth variables (ttal number f nests sampled, N= 2: May-PR = 4, June- PR= 3; May-PA=4, June-PA=4; May-NR= 4, June- NR = 1). The rder in which chicks frm each nest manipulatin were weighed was cunter-balanced with respect t time f day. Chicks frm each treatment grup were weighed and measured between 6-93 EST n 1 6, 9, 12, 15 and 18 days psthatching. Mass was recrded t the nearest gram using a Pesla scale, while length f the tarsus, 9th primary and 9th primary's sheath were measured t the nearest mm with a ruler. Cmparisn f age-specific mass acrss treatment (3 levels) and mnth f hatching (2 levels) was carried ut using data frm chicks aged 6, 12 and 18 days pst-hatching

4 176 via a three-way analysis f variance with repeated measures. A cmparisn f grwth parameters fr mass was made using a tw-way analysis f variance, with treatment (3 levels) and mnth (2 levels) as factrs. Average grwth curves fr each brd were iteratively calculated using all measurement data frm individuals by a three parameter lgistic mdel, W~ ~ A/{1 + exp[--k(t--tl)]}, where Wt is the mass f a chick at age t, A is the asympttic mass f a chick, K is the grwth cnstant and h is the mass f a chick at the inflectin f its grwth curve (NLIN prgram f Statistical Analysis System, 1985). Ttal 9th primary length and the rati f feather sheath and ttal 9th primary length were used as indices f plumage develpment. Linear measurements f chicks aged 9, I2, 15 and 18 days pst-hatching were analyzed using a tw-way analysis f variance with repeated measures. Bld analysis Because O. sylviarum is a bld feeding ectparasite, heavy parasitism lad may be reflected in affected chicks as decreased xygen transprt capabilities (bld hemglbin cntent r reduced red cell cunt). A subset f brds f 4 were cllected frm the field at l, 7, 13 and 19 days psthatching fr each f the nest treatment levels, and transprted t the labratry in insulated styrfam cntainers. The number f brds sampled fr the treatment, PR, was 11 ; with 2, 4, 3 and 2 brds taken at each f the sampling intervals, respectively. The number f brds sampled fr PA and NR treatments was 11 and 9, respectively; with 2, 3, 4 and 2 and 2, 2, 3, and 2 brds taken at each sampling interval, respectively. At the labratry bld samples were drawn frm the brachial vein fr analysis. A sample f 33 ~l f bld was cllected in heparinized capillary tubes and deep frzen (-5 ~ C) fr subsequent analysis. Ttal bld hemglbin was estimated clrimetrically by cmbining thawed whle bld (2 gl) with 5. ml Drabkin's slutin (1.25 g f sdium bicarbnate:ptassium ferricyanide :ptassium cyanide, 1: 2: 5 disslved in I l HzO and.5 ml 23-1auryl ether, 3% w/v), and allwing the mixture t stand at 23 ~ C fr 15 min. Samples were centrifuged at 3 rpm fr 2 rain t precipitate ut nuclei, and then decanted t spectrphtmeter tubes and cmpared t a Drabkin's reagent blank using a Spec-2 read at 54 nm. Hemglbin cntent f the bld was evaluated by cmparing absrbance readings t a cyanmethyhemglbin standard (human). A secnd bld sample was cllected in 75 mm micrcapillary tubes and spun at 25 rpm n a clinical centrifuge fr 15 min. Cmparisn f packed cell vlumes (vlume f cells/ttal bld vlume) are a reflectin f the number f bld cells per unit vlume, and as such are a crude indicatr f relative xygen transprt capacities f chicks. Chicks were euthanized and deep frzen (-5 ~ C) fr subsequent carcass analysis, the results f which are t be reprted elsewhere. The effects f treatment (3 levels) and age (4 levels) n bld hemglbin values were determined using a tw-way analysis f cvariance, with packed cell vlume as the cvariate. Labratry emergence tests During 1985, nests were brught int the labratry in an effrt t mre clsely mnitr the effects green plant mate- rial had n the emergence behavir f mites. Nest bxes were prepared and mnitred as they were during the 1984 field seasn. At 5 days pst-hatching, nests were cllected (N = 12), placed in ne galln 'zip-lck' plastic bags, brught t the labratry and randmly assigned t ne f three grups (N=4). After thrugh mixing f the matrix all nests were divided int halves and the halves were placed int separate 'zip-lck' bags. One half was designated as the untreated cntrl. The remaining, crrespnding, half was randmly assigned t ne f three plant treatment grups. Thse nests were mixed with 5 g f fresh green leaves f either wild carrt, fleabane r garlic mustard Alliaria fficinalis. In this way each nest served as its wn cntrl fr the number f mites emerging t feed. Feeding instars were cllected as they emerged frm the nest matrix. Mites were cllected fr 13 days after nests were first brught t the lab. The number f mites was quantified as described abve. N attempt was made t recver mites frm the nest matrix. Unless therwise stated, factrs in all analyses were fund t be hmgeneus. Pst-hc cmparisns f treatment levels were carried ut using the Neuman-Keuls technique with significance values reprted at P<_.5 (Winer 1971). Results Ornithnysus sylviarum ppulatins in the field The number f mites cllected frm nests increased ver the curse f develpment f chicks (F-37.4, df= 3,48, P <.1). This pattern did nt differ substantially between the first and secnd breeding attempts in May and June (F , df= 3,48, P =.212), thugh mre mites were cllected frm nests during June (F=72.59, df=l,48, P<.1). The bserved pattern f greater numbers f mites assciated with develpment f the yung and reuse f ld nests is cnsistent with data reprted fr ther passerines (Wasylik 1971). Manipulatin f the nests strngly affected the number f mites cllected (F=547.19, df--2,48, P<.1). Nests devid f any fresh plant material (PR) pssessed the largest mite ppulatins ver the curse f develpment f nestlings, while thse nests cntaining D. carta (PA) pssessed a significantly lwer number f mites. Nest-sites subjected t peridic replacement f nest material (NR) accumulated the fewest number f mites (Fig. l, F= 7.2, df= 6,48, P<.1). Clser inspectin f Fig. 1 indicates that additin f D. carta did nt affect the rate at which mite ppulatins increased relative t the plant remval treatment. Such a pattern suggests that survivrship and fecundity f mites were similar in these tw treatments. The lwer number f mites in the plant additin treatment may have resulted frm a delay in clnizatin r reprductin f the funding ppulatin. Effect f nest treatments n chicks The impact f the nest manipulatins n the age-specific mass f starling nestlings was negligible (F=.17, df= 45, P=.842, Fig. 2). Because nest treatments significantly cvaried with the densities f mites, we cncluded that the fwl mite had n appreciable effect n the age-specific mass f nestlings. Similarly, the verall mdel fr nest manipula-

5 177 5"? O 22 ~2 Z J 4 PR PA NR iiil DAYS POST HATCHING Fig. 1. The mean number f O. sylviarum cllected frm nests during the curse f develpment f starling nestlings. The lettered cdes crrespnd t the type f nest manipulatin perfrmed (NR: nest replacement, PA: plant additin, PR: plant remval). The vertical bars crrespnd t ne standard errr. The data presented are cllapsed ver the May and June breeding attempts #t v r-4 >,.-.1 r,...) 121 r_) c~ ~ ~ ~ cs ~ ~q~ Q 8B ~ go2 s I I I [ BLOOD HEMOGLOBIN (g-dl) Fig. 3. The relatinship between bld hemglbin cntent and packed cell vlume (r =.987, P<.1, N= 123) a ~ as 2- MAY JUNE PN PA m NR DAYS POST HATCHING PH NSN PA ~ NR 1i li'i DAYS POST HATCHING Fig. 2. The mean age-specific mass f chicks reared during the May and June breeding attempts. The lettered cdes crrespnd t the type f nest manipulatin (NR: nest replacement, PA : plant additin, PR: plant remval). Vertical bars are ne standard errr tin and seasnal effects fr the mean asympttic mass f chicks, the grwth cnstant and inflectin f the grwth curve well all fund t be nn-significant (F= 1.1, df= 5,14, P=.42; F= 1.39, df= 5,14, P=.285; F= 1.51, df= 5,14, P =.249, respectively). Figure 2 shws that the age-specific mass f chicks reared during the first breeding attempt in May differed frm their cunterparts reared in June (F= 4.75, df= 2,15, P= :11). Chicks reared in May were heavier than thse reared in June (F=4.9, df=1,75, P=.32). This is the typical pattern bserved fr mst passerines (O'Cnnr 1984). A study cnducted by Pwlesland (1977) in New Zealand n the effects f O. bursa n the grwth f starling nestlings yielded similar results. Ectparasite lad, did nt affect the grwth and develpment f the plumage. The age-specific lengths f the 9th primary were similar (F=.98, df= 6,72, P =.442), revealing n verall treatment effect (F=1.9, df--2,24, P=.171). Analysis f the rati f sheath length t ttal 9th primary length, as an index f maturatin f the plumage, yielded similar results. The rate at which feathers emerged frm the feather sheaths was similar fr all treatments (F= 1.5, df= 6,72, P=.42) and shwed n verall treatment effects (F=2.72, df=2,24, P=.86), thugh there was a trend fr chicks frm the plant remval treatments t retain their feathers in the sheath fr a slightly lnger perid. Because fwl mites are hematphagus we cnsidered the pssibility that parasite lad might affect the bld characteristics f nestlings. The rati f the vlume f red bld cells t plasma vlume (packed cell vlume) is a cnservative physilgical character. Even under dehydratin chicks maintain bld vlume t preserve smtic balance. Thus PCV, represents a gd relative index f the crude xygen carrying capacity f chicks (Fig. 3). Overall, nestlings reared within each f the three nest treatments exhibited differences in levels f bld hemglbin (F= 16.25, df = 2,11, P <.1), even after remving the cvariate effect f PCV (F=142.51, df=l,ll, P<.1, regressin apprach). Chicks reared in nests where D. carta was incrprated int the nest matrix, and chicks reared in bxes where nests were systematically replaced, maintained higher levels f age-specific bld hemglbin than thse chicks reared in the plant remval treatments (Fig. 4). We infer that the high mite densities assciated with the plant remval treatments were sufficient t depress the xygen carrying ptential f nestlings. Effectiveness f plants against O. sylviarum in the labratry The efficacy f treating nest material with green leaves was tested under mre cntrlled cnditins in the labratry. Emergence f feeding instars frm nest material was mnitred in nests where half the matrix was mixed with green leaves, while the emergence frm the remaining untreated matrix was mnitred as a cntrl. There was n difference in emergence f feeding instars frm nest material treated with the plant A. fficinalis relative t cntrls (Fig. 5A;

6 i, 2i v z i~- 1- i PR ~ PA II NR DAYS POST HATCHING Fig. 4. The mean hemglbin cntent f nestlings expsed t the three nest manipulatins thrughut the curse f their develpment. Treatments were NR: nest replacement, PA: plant additin, PR: plant remval. Vertical bars are ne standard A errr 25 [:Z] CONTROL 2 1 ALLDU~IA m 15 i 5 i 25 ~ ZS ZOO '~ 15 lo 5 _ PROTONY~aH 1 PROTONY~aH itdult 2 ~ [:=:3 1 DAUCUS CONTRO i,i. 1 i PROTONY~H [Z~ CONTROL 1 ERIGERON ADULT Fig. 5. The mean number f mites emerging frm nest material fr the halves frm each f a set f nests treated with green plant material (slid bars). The pen bars represent mean emergence f mites frm the untreated, cntrl prtin f nests. Vertical bars represent ne standard errr F=.1, df= 1,18, P=.924). Clark and Masn (1985) had previusly shwn that A. fficinalis was cmmnly available during nest cnstructin, but was statistically underutilized by starlings building nests. In cntrast, bth D. earta and E. philadelphieus were incrprated int nests in quantities greater than their prprtinal availability in the habitat. Inspectin f Fig. 5 B-C illustrates bth D. carta and E. philadelphicus significantly decreased the emergence f feeding instars f mites in the halves f nests treated with plants relative t the untreated cntrls (F= 6.5, df= B C 1,18, P=.23 and F= 6.8, df= 1,18, P=.24, respectively). Discussin Accrding t the nest prtectin hypthesis, plant chemicals fund in nest material act t reduce pathgen and ectparasite ppulatins nrmally fund in nest envirnments. If the use f green plants by birds is t be cnsidered adaptive several cnditins must hld true: (1) parasites and/r pathgens must, directly r indirectly, affect survival and/r fecundity f the breeding adults, (2) plants must in sme way decrease the ptential detrimenta! effects f parasites and pathgens, and (3) nest-building behavir must have a genetic basis and birds must preferentially select plant material that reduces parasite and pathgen lad. In additin, birds shuld pssess sensry capabilities that permit discriminatin amng plants. D ectparasites affect survival and reprductin? There is ample evidence t indicate that ectparasites influence fecundity and mrtality in birds. Our data demnstrate that large infestatins by mites result in reduced xygen carrying capacity f chicks. The extent t which lwer hemglbin levels influence pst-fledging survival in starlings remains t be determined, ther studies indicate that intensive feeding by O. sylviarum n chicks and adults f ther species can result in anemia r death f the hst (Hfstad et al. 1984). In the dmestic fwl Gallus spp, egg prductin is impaired by 5 t 15 percent as a cnsequence f mite infestatin (DeVaney 1979). Amng swallws and thrashers, parasitism by the hematphagus fly Prtealliphra can result in pr grwth f chicks (Mss and Camin 1971 ; Arendt 1985a; Brwn and Brwn 1986; Shields and Crk 1987). Furthermre, high infestatins f ectparasites disrupt breeding effrts by frcing adults t abandn nests (Feare 1976; Duffy 1983). In the case f bird species which typically reuse ld nests, large numbers f ectparasites frce adults t delay breeding until new nests are built (Brwn and Brwn 1986; Barclay 1988). Such delays can influence the prbability f successful secnd breeding effrts (Barclay 1988). Indirectly, hematphagus ectparasites can act as vectrs fr avian chlera, fwl px, viral rnithsis and Newcastle disease, t name but a few dcumented avian pathgens (Hfstad et al. 1984). Can plants effectively regulate ectparasite ppulatins? There is ample evidence t indicate that metablites cntained within plants can act as txins t arthrpds and pathgens (Frear 1948; Jacbsn 1954, 1975; Jacbsn and Crsby 1971). Our previus wrk in the labratry demnstrated that plants preferred by starlings effectively inhibited bacterial grwth and develpment f the hematphagus luse Menacanthus (Clark and Masn 1985). Hwever, the same study als shwed that D. carta, as well as ther plants, had n direct effect n survival f mites, even when the mites were allwed direct cntact with the plants. Paradxically, ur present data suggest that D. carta effectively reduces mite ppulatins in the field. We cnclude that it is unlikely that ppulatin cntrl f mites by D. carta is affected by lethal dses f txic cmpunds. Rather, the significantly lwer number f mites fund in

7 179 nests prtected by plants was presumably a result f a delay in the initial clnizatin f the nest by mites via repellency, inhibitin f vipsitin, r disruptin f the physilgical readiness f females t lay eggs. Because the lg number f mites was linearly related t the age f brds (Fig. 1), we were able t prject, via regressin, the number f mites required t clnize the nests. Nests nt prtected by plants (PR) needed apprximately 1 mites as clnists in rder t yield the ppulatin bserved. Mites clnize nests by leaving the bdies f already infested adults thrughut nest-building and incubatin (Pwlesland 1978). Infestatins f 5 t 2 mites are well within the range reprted t exist n adults (Byd 1951 ; DeVaney et al. 198). Alternatively, thse nests prtected by D. carta (PA) needed nly reduce the clnizing ppulatin size t 1-3 mites t yield bserved ppulatin levels. D. carta cntains the sterid, /~-sitsterl (Duke 1985), which is effective as a repellent and vipsitin inhibitr fr mites (Ambasta 198). Cincidentally,/~-sitsterl is als fund in the leaves f the margsa tree Azadirachta indica. Margsa leaves are used in nest-building by huse sparrws Passer dmesticus, much as D. carta is used by starlings (Sengupta 1981). The higher reslutin f ur labratry versus ur field data n the plant-mite interactin suggest still ther bilgical cntrls may be exerted n mites. Fr example, the time curse f emergence f mites frm the D. carta treatment reveals an interesting pattern (Fig. 6). Nt nly was there a differential emergence f mites frm cntrl bags and bags treated with D. carta, but als there was a delay in mlt assciated with treatment by plants. The majr effect f adding either D. carta r E. philadelphicus t nest.material was t decrease the number f mites emerging. Because psitive getaxis is ne f the principal behavirs assciated with feeding by mites, ne plausible interpretatin is that these plants cntain substances that act t disrupt feeding and behavirs assciated with feeding. Such cmpunds may stimulate inhibitry sensry prcesses which in turn suppress apprpriate rientatin behavir (Mustaparta 1984; Stadler 1984). Many plants cntain arthrpd juvenile hrmne analgs which can delay the transitin frm ne develpmental stage t the next (Slama 1979). The emergence f prtlymph-1 mites (thse never having fed) supprts the ntin that D. carta leaves may cntain cmpunds that delay the mlt f mites (Fig. 6). After tw days, many f the nn-feeding larva frm the cntrl half f nests mlted t the first feeding prtnymph stage, yielding a pulse f mites t be cllected as they emerged frm the nest matrix. Thereafter, recruitment t the first feeding prtnymph stage decreased and a nrmal expnential decay is emergence was bserved (as a cnsequence f remval sampling). Hwever, the emergence pulse attributable t recruitment in the halves f nests treated with D. carta was delayed by tw days. N shifts in the emergence f prtnymph-2 frms was expected because n mlt was invlved. The small numbers f adult stages emerging may have bscured any delays attributable t recruitment frm the deutnymph stage. Regardless f the mechanisms invlved, any delay r decrease in emergence f mites has tw cnsequences: 1) the per capita daily bld lss f chicks is decreased, and 2) withut hemlymph, mites wuld nt pssess sufficient energy t reprduce. The latter wuld delay the nset f i PROTORTy~PH 1 t i PROTONYMPH ADULT v~ CONTROL DAUCUS DAYS POST COLLECTION Fig. 6. The time curse f the emergence f mites frm nests treated with D. carta (slid bars) and their crrespnding untreated cntrls (pen bars). The vertical bars depict ne standard errr ppulatin grwth f mites until the agents respnsible decmpsed. The selectin f green plant material All plants cntain secndary metablites which functin as the plants' defense against disease and herbivry. Therefre, a randm selectin f green plants by birds might be sufficient t chemically prtect nests against their wn suite f pathgens and parasites. Hwever, in an earlier study we were able t demnstrate that starlings selected sme species f plants in quantities greater than their prprtinal availability in the habitat (Clark and Masn 1985). These preferred plants were mre likely t pssess antibacterial prperties relative t a randm subset f plants generally nt preferred by starlings. Our present data are cnsistent with this interpretatin. Of the species tested nly thse species preferred by starlings limited the emergence f mites. The species f green plants starlings prefer prbably change with lcale. Birds breeding in England and Ohi seem t shw preferences fr sme plants, but the species invlved are nt necessarily the same as thse preferred in Pennsylvania (P.W. Greg-Smith and R.A. Dlbeer, persnal cmmunicatins). We suspect that starlings use chemical, in additin t visual, cues t differentiate amng plants that act as pesticides. Visual cues, such as leaf shape r clr, may crrelate less well with bilgically active prperties f plants than d the emitted vlatiles. While there

8 18 are charactistic chemical differences that d crrelate with mrphlgical traits, there is still sufficient chemical variatin between individuals t suggest that visual discriminatin is an inadequate basis fr selectin f bilgically active leaves (Parks 1974). In ur earlier study we were unable t detect any mrphlgical similarity amng preferred plants. Hwever, we did bserve by gas chrmatgraphic analysis f vlatiles that plants preferred by starlings cntained mre cmpunds at higher cncentratins than a randm subset f plants drawn frm the envirnment (Clark and Masn 1985). At the present, we find it unlikely that starlings discriminate amng plants based upn specific vlatile cues attributable t specific bilgical activity. Rather, we suggest that starlings use general vlatility and chemical cmplexity as ne means f increasing chances f encuntering bilgically active plant material. Starlings pssess the requisite anatmical, physilgical and behaviral capacity t perceive and discriminate between plant drs using lfactin (Clark and Masn 1987). Whether they use this capability in the field as a basis fr chsing plants remains t be determined. Acknwledgments. Research supprt was prvided by a grant frm the Natinal Gegraphic Sciety, We thank K Heinzel fr assistance in the field. CA Smeraski was especially helpful fr assistance in washing nest material and srting instars f mites. References Ambasta SP (198) N title. Sci Rep 17:21 Arendt WJ (1985a) Philrnis ectparasitism f pearly-eyed thrashers. I. Impact n grwth and develpment f nestlings. Auk 12:27-28 Arendt WJ (1985b) Philrnis ectparasitism f pearly-eyed thrashers. II. Effects n adults and reprductin. Auk 12: Barclay RMR (1988) Variatin in the csts, benefits, and frequency f nest reuse by barn swallws (Hirund rustica). Auk 15:53-6 Byd EM (1951) A survey f the parasitism f the Starling Sturnus vulgaris L. in Nrth America. J Parasitl 37 : Brawn JD, Balda RP (1988) Ppulatin bilgy f cavity nesters in nrthern Arizna: d nest sites limit breeding densities? Cndr 9:61-71 Brwn CR, Brwn MB (1986) Ectparasitism as a cst f clniality in cliff swallws (Hirund pyrrhnta). Eclgy 67: Brush T (1983) Cavity use by secndary cavity-nesting birds and respnse t manipulatins. Cndr 85: Clark L, Masn JR (1985) Use f nest material as insecticidal arid anti-pathgenic agents by the Eurpean Starling. Oeclgia (Berlin) 67: Clark L, Masn JR (1987) Olfactry discriminatin f plant vlatiles by the Eurpean starling. Anita Behav 35: DeVaney JA (1979) The effects f the nrthern fwl mite, Ornithnyssus sylviarum, n egg prductin and bdy weight f caged white leghrn hens. Pult Sci 58: DeVaney JA, Quisenberry JH, Dran BH, Bradley JW (198) Dispersal f the nrthern fwl mite, Ornithnyssus sylviarum (Canestrini and Fanzag) and the chicken bdy luse, Menacanthus stramineus (Nitzch), amng thirty strains f egg-type hens in a caged laying huse. Pult Sci 59: Duffy DC (1983) The eclgy f tick parasitism n densely nesting Peruvian seabirds. Eclgy 64: Duke JA (1985) Handbk f medicinal herbs. CRC Press, Bca Ratn Feare CJ (1976) Desertin and abnrmal develpment in a clny f sty terns Sterna fuscata infested by virus-infected ticks. Ibis 118: Frear DEH (1948) A catalgue f insecticides and fungicides. VII. Chemical fungicides and plant insecticides. Chrnica Btanica C, Waltham, Mass Hfstad MS, Barnes H J, Calnek BW, Reid WM, Yder HW (1984) Diseases f pultry, 8th edn. Iwa State University Press, Ames, Iwa Hgstad O (1975) Quantitative relatins between hle-nesting and pen-nesting species Within a passerine breeding cmmunity. Nrw J Zl 23: Hgland JC, Sherman PW (1976) Advantages f Bank Swallw (Riparia riparia) clniality. Ecl Mngr 46 : Jacbsn M (1954) Insecticides frm plants: a review f the literature, USDA Handbk N 154, Washingtn, DC Jacbsn M (1975) Insecticides frm plants: A review f the literature, USDA Handbk N 461, Washingtn, DC Jacbsn M, Crsby DG (1971) Naturally ccurring insecticides. Dekker, New Yrk, Basel Jhnstn RF, Hardy JW (1962) Behavir f the Purple Martin. Wilsn Bull 74: Lye JE (1985) The life histry and eclgy f the cliff swallw bug, Oeciacus vicarius (Hemiptera: Cimicidae). Cab ORSTOM Set Entml Med Parasitl 23: Mss WW, Camin JH (1971) Nest parasitism, prductivity and clutch size in purple martins. Science 168 : 1~13 Mustaparta H (1984) Olfactin. In: Bell WJ, Carde RT (eds) Chemical eclgy f insects. Chapman and Hall, Lndn, pp O'Cnnr RJ (1984) The grwth and develpment f birds. Wiley, Chichester Parks CR (1974) Chemical evidence. In: Radfrd AE, Dickisn WC, Massey JR, Bell CR (eds) Vascular plant systematics. Harper and Rw, New Yrk, pp Pwlesland RG (1977) Effects f the haematphagus mite Ornithnyssus bursa n nestling Starlings in New Zealand. NZJ Zl 4:85-94 Pwlesland RG (1978) Behavir f the haematphagus mite rnithnyssus bursa in Starling nest bxes in New Zealand. NZJ Zl 5 : Rthschild M, Clay T (1957) Fleas, flukes and cucks. Cllins, Lndn SAS Institute Inc (1985) SAS user's guide: statistics. Cary, NC Sengupta S (1981) Adaptive significance f the use f margsa leaves in nests f the huse sparrw Passer dmesticus. Emu 81 : Shields WM, Crk JR (1987) Barn swallw clniality: a net cst fr grup breeding in the Adirndacks? Eclgy 68: Slama K (1979) Insect hrmnes and antihrmnes in plants. In: Rsenthal GA, Janzen DH (eds) Herbivres. Their interactin with secndary plant metablites. Academic Press, New Yrk, pp Stadler E (1984) Cntact chemreceptin. In: Bell WJ, Carde RT (eds) Chemical eclgy f insects. Chapman and Hall, Lndn, pp3 36 Stner D (1936) Studies n the Bank Swallw Riparia riparia (Linueaus) in the Oneida Lake Regin. Rsevelt Wildl Ann 4: Sykes RK, Chamberlain RW (1954) Labratry bservatins n three species f bird mites. J Parasitl 4: Tenvu R, Lemmetyiuen R (197) On the breeding eclgy f the Starling Sturnus vulgaris in the Archipelag f suthwestern Finland. Ornis Fenn 47: Wasylik A (1971) Nest types and abundance f mites. Ekl Pl 19 : Wimberger PH (1984) The use f green plant material in bird nests t avid eetparasites. Auk 11: Winer BJ (1971) Statistical principles in experimental design. McGraw-Hill, New Yrk Received January 11, 1988