31 jul}' 1978 JOURNL\L OF TliE KANSAS ENTOMOLOGICAL SOCIETY 51(3), 1978, pp. 432-440

ECOLOGY OF NELTfJ"AtllUS TEXANUS (COI..~EOPTERA: BRUCHIDAE) IN THE SEEDS OF (RI-IAf\,fNACEAE)

Clarence Dan Johnson

Abstr(Jct. Ncllurni.us lexnnus spends its larval ·]ife in sr.cds of Condalia correllii, C. globosa var. pulJescens, C. hookeri var. hookerl, C. spathula.la, and C. tvar11ockii var. kearneyana. Eggs of N. texanus are glued to the sur­ face of fruits of Condalia. The larvae enter the fruit, feed jnside a seed, and pupate in it. Only 1.4% of the fruits collected had adults etnerge froJn them although more had b1uclrid eggs on their surfaces. The other two species of l\7elfu1n.ius, N. gibbithorax and N. arizonensi.s, feed only in seeds of the legun1e genus Prosopis. Some reasons for the unusual feeding habits of N. texanus are discussed. The reeent separation of the genera Conclalia and Ziziphus · is supported because the hn1chids that feed in Conclali.a and Ziziphus are different. N. texanus bas a distribution fron1 Texas to Cali­ fornia into northc111 Mexico.

Because n1ost repo1t~ of the host of bruchid beetles arc of then1 feeding in seeds of the fanJily Legun1inosae, any report fron1 seeds of an­ other fatnily js of interest. TJ1is is pn.rticularly true 'hrith bruchicls in the genus Neltumius because both N~ arizon.ensi.s (Schaeffer) and N. gibbithorax (Schaeffer) are knO'\IVn to feed only il1 seeds of the legun1e genus Prosopis. When N. texanus (Schaeffer) \vas first reported to attack seeds of a Con.dalia by Kingsolver (1964) it \vas rcgal'dcd witl1 ~kepticisn1 becnuse of tlle known preferences l1y the other two species for Prosopis. 1\earings by l.~. J. Dott.imer and I have substautiatcd that N. texanus does feetl in seeds of several species of Condalia. Because of the uniqueness of this host­ insect interaction, I describe the ecology and behavior of N. texat1us below.

l\:fethods The methods of Johnson (1970) were used in nccnmulating the rearing data presented in this paper. In the text of this paper, C. D. Johnson is abbreviated CDJ and L. J. 13ottimer as LJB. The scannjng electron Jnicro­ seopc photographs were taken with an A1vtR 1000 SEM.

Results The host. The most recent revision of Co-nclaUa was by Johnston (1962) in '~7 hich be treated sorne 18 Nortl1 and South Atne1ican species. These pl~nts are spiny or rarely sn1all \vith stnall, d1upaceous fruits

.. VOLUME 51, NUMBER .3 433

Table 1. Classification of Condalia spp. according to Kearney and Peebles (1969) and Johnston (1962).

C. globosa = C. globosa var. pubescens C. spathulata = C. warnockii var. kearneuana C. mexicana ::::: C. correllii and have a disbibution from to Texas south to Argentina. John­ stem's classification differs from that of Kearney and Peebles (1969) and is followed in this paper (Table 1). The species in Arizona and northern Mexico are loca11y abundant hut not common, and usually grow at elevations above 3,000 feet. The three species in Arizona grow in the ­ scrub (both Arizona Upland and Lower Colorado Subdivisions), Chihuahuon Desertscrub, and Plains and Desert Grassland of Lowe and Brown (1973). Condalia globosa var. rml;escens (Table 3) matures fruit at least during April, C. correWi matures fruits dming Octoher and N ovcmher, and Condalia warnockU var. kearne!f(Jn.a has mature fruit on plants at different localities from at least August to November. According to Takhtajan (1969) the Rhanmaceac arc closely related to the Vitaceae. This is of interest to me because species of both families have their seeds attacked by species of bruchids. Johnson and Kingsolver (1975) reported that the Arizona wild grape (Vi.tis arizonica) is infested by i\m/,/ycerns oUis (Schaeffer) and Forister and Johnson (1971) described the bionomics of Acanthoscelides prosopoides (Schaeffer) which feeds in seeds of Ziziphus obtusifolia (Rhamnaccae). BarcetUI, Condalia and Rhamnus are other rhamnaceous genera that have been reported to have their fruits infested by bruchids. The infestation.-Of the 22 samples of fruits of C. globosa, C. warnockii, and C. correllii collected in Arizona and northern ~1exico, only nine samples yielded adult bruchids (Table 2). Most of the 22 samples had at least some mature fruits in them and most had bruchid eggs attached to the fruits. Several cultures of predominantly mature fruits had several bruchid eggs attached to them but no adults emerged. Seeds of C. globosa collected in April 1975 yielded many adults but mature fruits of C. warnockii and C. correllii with eggs on them collected in August 1975 yielded no adults. Fruits of the latter two species coUected in October and November 1976 yielded adult N. texanus. Perhaps laboratory conditions were not suitable for larval development for those collected during August. More likely the seeds require a maturation period on the before they become suitable nourishment for bruchid larvae. I have examined herbarium specimens of Condalia spp. collected in Arizona and California. Several of the specimens of all three species from Arizona had either eggs and/or exit holes of bruchids in their fruits. These 434 JOURNAL OF THE KANSAS ENTOMOLOGICAL SOCIETY

Table 2. Emergence data for Neltumius texanus in seeds of Condali.o spp.

Culture# Date emerged by #Emerged

Condalia correllii #107-76 1-10-77 2 1-12-77 1 Il-28-77 1 C. globosa var, pubescens #4-75 V-16-75 2 V-18-75 10 V-20-75 9 V-21-75 2 V-23-75 10 V-25-75 10 V1- 2-75 13 VI-I3-75 2 VII- 9-75 3 IX-1 1-75 2 #5-75 V-18-75 1 V-20-75 I #G-75 V-16-75 2 V-18-75 3 V-20-75 J V-21-75 1 V-23-75 2 VI- 2-75 1 #7-75 V- 5-75 I V- 8-75 2 V- 9-75 4 V-12-75 5 V-I3-75 11 V-14-75 7 V-16-75 7 V-18-75 22 V-20-75 l9 V-2l-75 9 V-23-75 12 V-25-75 18 VI- 2-75 I2 Vl-13-75 2 VII- 9-75 I VIII-15-75 9 1-29-76 3 #8-75 V-20-75 1 V-2.5-75 1 1X-11-75 1 C. wamockii var. kearneyana #33-76 XI- I-76 2 XII- 8-76 I XII-22-70 1 III- 7-77 1 VOLUME 51, NUMBER 3 435

Table 2. Continued.

Culture# Date emerged by # Emerged #34-76 X-25-76 2 XI-12-76 3 XII- 8-76 I XII-22-76 I UI-22-77 2 lV- 5-77 3 #.35-76 XI-12-76 1 XII-22-76 I C. hookeri var. l!ookeri 89b XII- 9-47 I 9lm UI-15-49 3 124£ VI- 6-65 6 C. spatlmlata 96k VI- 4-56 I were collected within the range encompassed by the plants listed in Table 3. Bruchid eggs and exit holes were observed in fruits of C. globosa col­ lected north of Ogilby, near Picacho Peak, Midway Well, and the Chocolate Monotalns, all ln eastern Imperial Cmmty, California. Tho presence of bruchicls in these fruits is to be expected because the localities are across the Colorado River from C. globosa l1arboring brnchids in Arizona. Even though few N. texanus have been co1lected in Arizona and none in Cali­ fornia, because of damage to Condalia seeds we can safely say that this species ranges into California. It most likely ranges into and , Mexico, but, to my knowledge, it has never been co11ected in these states. Life htstory.-Sixtecn seeds of Condalia globosa (#4-75), 11 seeds of C. warnackii (#34-76) and 14 seeds of C. correllit (#107-76) were soaked in an ammo11ia solution and then dissected to learn man~ about the habits of N. texanus. All of the fruits dissected had only one hard, bony seed in them. The watcrsoaked pu1p of C. correllii and C. globosa was about 1 mm thick hut the pulp of C. warnockit was about 2 mm thick. When the pulp was removed from the seeds, those of C. correllii were about 5.3 X 3.5 X 3.2 mm, those of C. warnockii about 4.1 X 2.1 X 2.7 mm, and those of C. globosa 3.5 X 3.0 X 2.5 mm. Eggs of N. texanus are glued to the outside of the fruits (Figs. 1, 3, 5) or beneath the calyx by means of an adhesive beneath the egg and a sur­ rounding flange (Figs. 7, 8). The larvae bnrrow directly through the por­ tion of the egg attached to the fruit, then directly through the pulp and into the seed. Sometimes several eggs may partially cover each other (Figs. 3a, 5a, 8). In some species of bruchids this is known to be a protection against all eggs being destroyed hy egg parasites lwcause only the eggs on top arc parasitized. The egg chotia usually are filled with a whitish rnatprial ap- 436 JOURNAL OF THE KANSAS ENTOMOLOGICAL SOCIETY

Table 3. Host plants and collection localities for Neltumius texanus.

Old Record: Kingsolver, 1964: Condalia obovata Hooker? New Record.'lf: Condalia correllii M. C. Johnston: Arizona. Cochise Co.: en. 4,400', 8 mi E Dougla,, 6-XI-76 (CD) #107-76). C. gfobosa var. pubescens I. M. John.slon: Arizona. Yuma Co.: ncar Horse Tanks, Castle Dome l\Hs., 25-IV-75 {CDJ #4-75). Pima Co.: 10 mi N Lukeville, 26-IV-75 (CD) #5-75); 3 mi S Why, 26-IV-75 (CD) #6-75); 2 mi S Why, 26-IV-75 (CDJ #7-75); 10 mi W Quijotoa, 26-IV-75 (CD) #8-75). C. lwokeri var. lwokeri 1-.·1. C. Johnston: Texas. Cameron Co.: Palm Grove, Browns­ ville, 19-IX-47 (LJB #89b); Hidalgo Co.: Edinburg, 28-IX-64 (LJB #124£); Uvalde Co.: SW Uvalde County, ll-XI-48 (LJB #9lm); Uvalde, 15-III-49 (L. J. Bottimer coli.}. C. spathulata A. Gray: Texas. Webb Co.: La Mesa Ranch, 11-X-55 (LJB #96k). C. wamockii var. kearneyana M. C. Johnston: Arizona. Pinal Co.: 2.5 mi N Oracle Junction, 13-X-76 (CDJ #33-76); Pima Co.: ca. 3,500', ca. 4 mi \V Colossal Cave, 14-X-76 (CDJ #34-76); ca. 3,000', 1 mi W Headquarters, Saguaro Nat'l. Mon., 14-X-76 (CD) #35-76).

parently produced by the larvae as they enter the fruit (Figs. 1, 3, 5, 7, 8). More than one ]arva entered several of the seeds hut none or only one adult emerged from these. As in other bruchids, cannibalism apparently occurs in N. lexanu.s. The larvae feed inside the seed, and apparently molt several times as small amounts of frass and cast skins remain behind inside the seeds. From 0.5 to 0.9 of the seed contents are consumed by the larvae. The amount consumed is apparently dependent on the size of the seed (i.e., the larvae use a greater amount of a small seed). The larvae pupate inside the seed but do not spin cocoons. The adults emerge through typical round exit holes (Figs. 2, 4, 6) which are usually located at the apex of the seeds but some emerged at the base near the peduncle in C. warnockii. The adults may feed on the flowers of the hosts as I have col­ lected them by sweeping plants in flower. Neltumius texanus did not have a high percent infestation in any culture (Table 4). Culture #7-75 had the highest infestation with 3.05%. The in­ festation for all cultures was 1.4%, a very low infestation rate when com· pared to other b1uchids and seeds (Janzen, 1969; Johnson, in preparation) although Johnson and Kingsolver (1975) reported a similar percentage in­ festation for Amblycerus vitis feeding in Vitis arizonica seeds. If the f1uits of the 13 samples that were collected and no bruchids reared from them are considered, then the infestation rate is a slightly lower 1.1 %. Seeds of C. globosa from western Arizona (Table 4, cultures 4-75, 6-75, 7-75) were slightly more heavily infested than seeds of the other two species from southeastern Adzona. There are too many variables (temporal biotic and abiotic factors, host variability, etc.) that could account for the differ­ ences in infestation to draw many conclusions regarding the low rates of VOLUME 51, NUMBER 3 437

Figs. 1-4. 1. Fruit of Condalia glo1Josa var. pubescem with egg of Ncltwnirts texmws on its surface; 2. Fruit of Condalia glollosa var. pubesccns with exit hole of Neltumius te:1:anu8 adult ncar apex; 3. FI"Uit of Cowlalia warnockii var. kearne!fana with two overlapping eggs (a) of Neltumius texanus on its surface; 4. Fruit of Condalia wanwckii var. kcarneyana wilh exit hole of Ncltwnim tcxanus adult near apex. infestation. Seeds of C. warnockii and C. correllii were collectec.l several times from the same and different localities but bruchids were reared from them only when they were collected in October and November, respectively. Again, perhaps seed maturity is necessary before proper development of larvae can occur. According to the small amount of data presented here, unless N. texanus breeds in seeds while they are on the ground, it is not likely that they have much effect on the population dynamics of Com/alia spp. Periods of peak emergence were shortly after the samples were collected for the two samples where there is a pattern (Table 1, #4-75, 7-75). It is possible that there is more than one generation per year but if these beetles continue to breed in seeds, they do not do so very effectively in the lab­ oratory. The data presented arc inadequate to draw conclusions at this time (Table 2), 438 JOURNAL OF TilE KANSAS ENTOMOLOGICAL SOCIETY

Figs. 5-8. 5. Fruit of Condalia correllii with three overlapping eggs (a) of Neltumius texanus on itS surface. Note two eggs on portion of another fmit in lower left hand comer; 6. Fruit of Condalia correllii with exit hole of Neltumitts texanus adult ncar apex; 7. Closeup of egg of Neltumius tcxmws on fruit of Condalia globosa var. pubescens. Note flange (b) of adhesive surrounding the egg; 8. Closeup of three eggs of Neltumius texantts on fruit of Condalia correllii. Note the eggs overlap each other and the flange (b) of adhesive surrounding the eggs.

Discussion A bruchid that feeds in the seeds of a plant in the family is unique, but a bntchid in the genus Neltumius that feeds only in seeds of a rhamnaceous plant is even more remarkable. This is so because of the two other species in the genus, Neltumius gihbit1wrax is known only from seeds of Prosopis pubescens and N. arizonen.sis only feeds in seeds of P. fuliflora. Because about 85% of bruchids are known to feed in seeds of the Leguminosae (Johnson, 1970), I speculate that species ancestral to today's Neltumius species originally fed in leguminous seeds. Pressures of host plant defense mechanisms and perhaps competition from other seed-eaters have forced them to move to a secondary host, Condalia. Neltumius gi1Jbithorax is limited in its distribution to California, Adzona, Utah, Nevada, and northern Mexico. N. arizonensi.s is limited to Arizona, Nevada and Texas and N. texanus to Arizona, California, Texas and northern Mexico. Thus, all three species have a similar distribution. The disbibu- VOLUME 51, NUMBER 3 439

Table 4. Percentage infestation of seeds of Condalia spp. by Neltumius texanus.

#culture #seeds # % #wasps 4-75 5,273 63 1.2 9 5-75 2,698 2 0.07 5 6-75 781 10 1.3 12 7-75 4,617 141 3.05 21 8-75 778 3 0.4 I 33-76 1,620 5 0.3 0 34-76 1,446 12 0.8 8 35-76 147 2 1.4 4 107-76 565 4 0.7 11 TOTAL 17,925 242 1.4 TOTALS C. globosa 14,147 219 1.5 C. wamockii 3,213 19 0.6 C. conellii 5G5 4 0.7

tion patterns of Conda7ia and Prosopis are also similar. I hypothesize that ancestors of N. texanus originaJly attacked seeds of a Prosopis species and due to competition moved into the unattacked seeds of Condalia where there was little or no competition. The similar disbibutions of their hosts would have made this host switch relatively easy because species of both plant genera are in close proximity to each other. Johnston (1962, 1964) placed Comlalia lycioides in Ziziphus (as Z. obtusifolia) on the basis of morphological characters. :t-.1y studies of host preferences support this separation because a very different bruchid, Acanthoscelides prosopoides, attacks the seeds of Z. obtusifolia. In addi­ tion, a new species similar to A. prosopoldcs hns hem n•nrcd from Zizlplws from South America by A. L. Teran, further ecological evidence that these two plant genera are separate and distinct.

Hymenoptera Most of the Hymenoptera reared from Condalia seeds were Chalcidoidea with the exception of four Braconidae reared from #8-75 (one specimen) and #34-76 (three specimens). Another braconid was reared from #113-76, a culture from which no bmchids emerged. At least three species of Chalcidoidca are represented among the other wasps with the great majority in the genus Bruchophagus, probably B. borealis Ashmead, a known para­ site of bruchids. Only cultures 6-75 and 7-75 show appreciable numhcrs of these parasites. One culture (#68-75, Condalia warnockii, from 5 miN Portal, Cochise Co., AZ, collected 11-VIII-75) yielded 47 specimens of this species but no bruchids. The 802 seeds of this culture were mature when 440 JOURNAL OF THE KANSAS ENTOMOLOGICAL SOCIETY collected and had some bruchid eggs on their surfaces. Normally at least some bruchids should have been reared from these seeds but probably parasitization completely annihilated the bruchid population. Parasitism, and perhaps competition if some of the chalcids arc seed­ eaters, may have a limiting effect on the number of bruchids reared from a given collection of seeds.

Acknowledgments M. Johnson, R. Johnson and J. Johnson assisted with collecting seeds while J. Floyd and E. Conroy helped with cultures in the laboratory. For this I am most grateful. I thank E. C. Becker and D. E. Bright, Canadian Na­ tional Collection of Insects, Ottawa, for making available the host records of L. J. Bottimer. I am also grateful to J. N. Grim, Northern Arizona Uni­ versity, for taking the SEM photographs. Partial support for collecting wns prqvided hy the Insect Identification and Beneficial Insect Introduction Institute, U.S. Department of Agriculture, under Grant 12-14-100-9970 (33).

Literature Cited Fori<>ter, G. W., and C. D. Johnson. 1971. Behavior and Ecology of Acantlwscelides prosopoides (Coleoptera: Bruchidae). Pan-Pac. Entomol. 47:224-234. Janzen, D. H. 1969. Seed-eaters versus seed size, number, toxicity and dispersal. Evolution 23:1-27. Johnson, C. D. 1970. Biosystematics of the Arizona, California, and Oregon species of the seed beetle genus Acanthoscelides Schilsky (Coleoptera: Bruchidae). Univ. Calif. Publ. Entomol. 59:1-116. Johnson, C. D., and J. M. Kingsolver. 1975. Ecology and redescription of the Arizona grape bruchid, Amhlucerus vitis (Coleoptera). Coleop. Bull. 29:.'321- 331. Johnston, M. C. 1062. Revision of Cm1dalia including Microrlwmnu.s (Rhamnaceae). Brittonia 14:332-368. 19G4. The species of Ziziplms indigenous to United States and Mexico. Amer. Jour. Bot. 50:1020-1027. Kearney, T. H., and H. I-I. Peebles. 1969. Arizona Flora. Univ. Calif. Press, Berkeley and Los Angeles. 1085 pp. Kingsolver, J. M. 1964. The genus Neltumius (Coleoptera: Bruchidae). Coleop. Bull. 18, 105-111. Lowe, C. H., Hnd D. E. Brown. 1973. The Natural Vegetation of Arizona. Arizona Resources Information System Cooperative Publ. No. 2, Sims Printing Co., Phoenix. 5.'3 pp. Takhtajan, A. 1969. Flowering plants, origin and dispersal. Smithsonian Inst. Press, Washington, D.C. 310 pp.

Department of Biological Sciences, Northern Arizona University, Flag­ staff, Arizona 86011. Received for publication 27 June 1977.