AND RELATED HOMOPTERA OF SHADSCALE: POSSIBLE AGENTS IN THE DIEOFF PROBLEMINTHEINTERMOUNTMN WEST C. Riley Nelson B. Austin Haws David L. Nelson

ABSTRACT document the incidence and distribution of these in the large stands of shadscale found in the Great Basin Results ofa survey of the Homopteran fauna associated and on the Colorado Plateau. Another important objective with shadscale (Atriplex confertifolia) are reported. Eight of this study was to examine the possible relationship be­ coccoid Homopteran species in eight genera and four tween population levels of several species of Coccoidea and families were found: Coccidae, Ceroplastes irregularis; the dieoff problem. Diaspididae, Aonidomytilis incisus; Ortheziidae, Orthezia annae; Pseudococcidae, Chorizococcus polyporus, Dis­ tichlicoccus salinus, Humococcus atriplicis, Phenacoccus MATERIALS AND METHODS solenopsis, and Puto atriplicis. Root-dwelling forms were From a general survey of the insects on shadscale sampled by digging more than 2,000 plants in the Great (Haws and others, these proceedings) mealybugs and scale Basin ofIdaho, Nevada, and Utah as well as the Colorado insects were found most often in the crown and upper root Plateau of eastern Utah. Relationships between the pres­ zone of the plants (fig. 1). In that study, a backhoe was ence ofcoccoids and the health of individual plants and used to dig a limited number of plants ofshadscale and sites are summarized. fourwing saltbush ( [Pursh] Nuttall) plants. No coccoids were encountered more than 15 em INTRODUCTION below ground. Careful examination of the aboveground portions of the plants revealed very few coccoids. There­ As emphasized throughout this symposium, large tracts fore, in this study we concentrated our observations in the of shadscale (Atriplex confertifolia [Torrey & Fremont]) area of the crown and upper roots of the plants. All plants are dying throughout the Intermountain West. A variety dug were sacrificed during sampling and therefore were of causes have been hypothesized for this dieoff, including unavailable for future sampling. plant decline caused by damage. During a general Uniform stands of shadscale near established roads survey of the insects associated with shadscale communi­ of the Great Basin and Colorado Plateau were considered ties many herbivores were found feeding on shadscale for sampling, and individual sites were chosen after onsite (Haws and others, these proceedings). Among these visits were made. Sites were located, whenever possible, in herbivores were large numbers of several species of scale stands of known ploidy (Stutz and Sanderson unpublished insects (Homoptera: Superfamily Coccoidea sensu Borror, data) for future reference and study. Once a sampling site Triplehorn, and Johnson 1989). was chosen, a line for the sampling transect was deter­ Since scale insects and mealybugs are important pests mined by sighting on a prominent landmark roughly per­ of ornamental shrubs and fruit trees (Little 1972), we pendicular to the road. To establish the basal point of the focused our attention toward an indepth survey of coccoid transect we walked the transect line to a point where im­ Homoptera on shadscale. Our prime purpose was to pact from the roadway appeared negligible, then paced an additional5 mas a buffer. At 10 points on the transect, each separated by 10m, we noted the condition of the shadscale plant nearest the point and ranked it on a scale Paper presented at the Symposium on Cheatgrass Invasion, Shrub Die­ Off, and Other Aspects of Shrub Biology and Management, Las Vegas, NV, from 0 to 3, a rank of 0 representing a dead plant, a rank April 5-7, 1989. of 1 having 1 to 33 percent of its branches living, 2 having C. Riley Nelson is the Tilton Fellow, Department of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, CA 34 to 66 percent living branches, and 3 having 67 to 100 94118. Present address: Brackenridge Field Laboratory and Department percent living branches. The conditions of these plants of Zoology, University of Texas, Austin, TX 78712. B. Austin Haws is were used in calculating a Site Condition Index (SCI) by Professor Emeritus, Department of Biology, Utah State University, Logan, UT 84322. David L. Nelson is Research Plant Pathologist, Intermountain taking the average of the ranked condition of the 10 plants Research Station, Forest Service, U.S. Department of Agriculture, Shrub along the transect. SCI values thus vary between 0 and 3. Sciences Laboratory, Provo, UT 84606.

This file was created by scanning the printed publication. 152 Errors identified by the software have been corrected; however, some errors may remain. paper bags or Riker mounts. All specimens were labeled with capture locality data including: State, county, local­ ity, date, collector(s) name, and a unique number (CRN number). We thank Dr. Raymond J. Gill of the California Department of Food and Agriculture for generic and spe­ cific identifications of coccoids found during this survey. Voucher specimens are deposited in the Range Insect Collection of the Department of Biology at Utah State University and some additional specimens are in Dr. Gill's collection.

Collection Sites Sampling sites were chosen at localities throughout the Intermountain West (fig. 2). In the following list of collec­ tion sites the first number is the locality number, which will be cited under the individual taxa in the systematic section and in the distribution table (table 1) that follow. The CRN numbers are field notebook numbers taken by C. R. Nelson (additional information for those sites with CRN numbers maybe obtained from C. R. Nelson). The Site Condition Index (SCI) is given in parentheses follow­ ing the locality and collector data for those sites where it was calculated. The coccoid fauna associated with each of these sites can be found under the sytematic treatments of the individual taxa. Sites from which no coccoids were found are indicated with an asterisk (*). 1. IDAHO: Cassia Co., 4 mi E of Malta, 26 April1988, T. Evans, B. A. Haws, & C. R. Nelson, CRN #5159. Figure 1-0rthezia annae on shad­ No SCI taken. scale. Utah, Tooele Co., locality 47. Scale bar= 10 em. 2. IDAHO: Cassia Co., 14 mi SSW Malta, Geothermal Road, 26 April 1988, T. Evans, B. A. Haws, & C. R. Nelson, CRN #5160. No SCI taken.

We define incidence to be the percent of plants harboring coccoids. Incidence can refer to the scale of either within or between sites. At each of the 10 points on the transect we dug four plants and examined them for coccoids. These plants included one each of conditions 1 through 3 and one juve­ • • nile plant (where possible). The plant of each of these •• categories nearest the transect point was sampled. The plant used in calculating the SCI was one of the four plants dug at each transect point. Plants were judged to be juvenile if their canopy was less than 15 em, their crown less than 1 em in diameter, and they bore no evi­ ~ .. dence of flowers or fruit. All juveniles sampled were of condition 3. The other three plants taken at the transect point were all nonjuvenile (by these criteria). Nearly all . :.. ?!( ., .. juvenile plants were erect with an evenly tapering tap­ • • root. The other plants sampled varied from prostrate •• • to erect with taproots of a variety of configurations. As each plant was dug, care was taken to gently remove the soil to minimize dislodging of coccoids. All coccoids present were identified to family onsite using the family classification scheme given in Borror and others (1976). Numbers of each family were then counted separately. Voucher specimens were preserved for specific identifica­ tion offsite. Nonsessile forms were placed in vials con­ Figure 2-Distribution of shadscale sites sampled taining 70 percent ethanol; sessile forms were placed in for Coccoidea, 1988-89.

153 Table 1-Distribution of Coccoidea by locality number listed in text. Orth-Orthezia annae, Chor-Chorizococcus poly- porus, Dist-Distichlicoccus salinus, Humo-Humococcus atriplicis, Phen-Phenococcus so/enopsis, Puto-Puto atriplicis, Cero-Ceroplastes iffegu/aris, and Aoni-Aonidomytilis incisus. Pseu-Pseudococcidae undetermined. Question marks (?) designate species unconfirmed by taxonomic authority

Locality No. Orth Chor Dlst Humo Ph en Puto Cero Aonl Pseudococcldae

1 Orth Pseu 2 Orth 3 Pseu 4 Aoni? Pseu 5 Pseu 6 Pseu 7 Orth 8 Aoni? Pseu 9 10 Orth Chor Ph en 11 Chor 12 Cere Aoni Pseu 13 Orth Cere 14 Orth 15 Orth Aoni 16 17 18 Orth Chor 19 Orth Humo 20 Chor Ph en 21 Chor Dist 22 Dist Ph en 23 Chor 24 Dist Ph en 25 Ph en 26 27 28 Ph en Puto 29 30 Chor Ph en 31 Pseu 32 Puto 33 34 Pseu 35 Puto 36 Ph en Puto 37 Chor Humo Ph en 38 Chor Puto 39 Orth Ph en 40 Chor Dist 41 Puto 42 Orth Pseu 43 44 Orth 45 46 47 Orth Ph en 48 Orth Pseu 49 Pseu 50 Orth Pseu 51 Pseu 52 Orth 53 Orth Ph en 54 Ph en 55 Pseu 56 Chor 57 Chor Dist 58 Orth Chor

154 3. IDAHO: Owyhee Co., 15 mi S Bruneau, Hwy 51, mile­ 21. UTAH: Grand Co., 1.5 mi N Crescent Junction, marker 57 (13 mi S junction Hwys 78 & 51), 26 July Thompson Pass Road, 9 June 1988, R. W. Baumann 1989, B. A. Haws & C. R. Nelson, CRN #5377. (2.7). & C. R. Nelson, CRN #5227. (2.1). 4. IDAHO: Owyhee Co., Murphy, Hwy 78, NW hillside 22. UTAH: Grand Co., 1.5 mi N Crescent Junction, above bridge, 26 July 1989, B. A. Haws & C. R. Nelson, Thompson Pass Road, 12 July 1988, B. A. Haws & CRN #5379. (2.4). C. R. Nelson, CRN #5264. (1.8). 5. NEVADA: Eureka Co., 10 mi S Eureka, Hwy 50, 23. UTAH: Grand Co., 52 mi E Green River, 2 mi N Duckwater turnoff(Fish Creek Road), Hwy 379,25 I-70 on road to Cottonwood Ranch, 25 May 1988, July 1989, B. A. Haws & C. R. Nelson, CRN #5376. C. R. Nelson, CRN #5206. No SCI taken. (1. 7). 24. UTAH: Grand Co., 52 mi E Green River, 2 mi N 6. NEVADA: Humboldt Co., Golconda, eastbound I-80 1-70 on road to Cottonwood Ranch, 9 June 1988, S of freeway near exit 194, 12 August 1988, J. K. & R. W. Baumann & C. R. Nelson, CRN #5228. (2.0). C. R. Nelson, CRN #5281. (1.2). 25. UTAH: Grand Co., 52 mi E Green River, 2 mi N 7. NEVADA: White Pine Co., 15 miSE Eureka, 1 miN 1-70 on road to Cottonwood Ranch, 12 July 1988, Hwy 50 on Hwy 892, milemarker 1, 25 July 1989, B. A Haws & C. R. Nelson, CRN #5265. (1.2). B. A. Haws & C. R. Nelson, CRN #5375. (1.8). 26. UTAH: Grand Co., Professor Valley, Hwy 128 8. OREGON: Malheur Co., 2 mi S Hwy 201/19 on Succor milemarker 17, 9 June 1988, R. W. Baumann & Creek Road, 27 July 1989, B. A Haws & C. R. Nelson, C. R. Nelson, CRN #5230. (1.6). * CRN #5380. (2.6). 27. UTAH: Grand Co., Professor Valley, Hwy 128 mile­ 9. OREGON: Malheur Co., 15 mi SW Vale on hillside marker 17, 12 July 1988, B. A. Haws & C. R. Nelson, road cut, Hwy 20, 27 July 1989, B. A. Haws & C. R. CRN #5266. (1.6). * Nelson, CRN #5384. (2.8). * 28. UTAH: Juab Co., Deep Creek Mountains, CCC 10. UTAH: Beaver Co., 8 miN Milford, Hwy 257 mile­ Campground Callao, 3 June 1988, C. R. Nelson, marker 8, 14 June 1988, C. R. Nelson, CRN #5244. CRN #5222. (1.8). (1.7). 29. UTAH: Millard Co., 1 mi W Intermountain Power 11. UTAH: Beaver Co., Wah Wah Valley, Hwy 21, 22 mi Project, S side of road, 4 May 1988, T. Evans & W Milford, 15 June 1988, C. R. Nelson, CRN #5247. B. A. Haws. No SCI taken.* (3.0). 30. UTAH: Millard Co., 11.7 mi W Leamington, Hwy 6 12. UTAH: Daggett Co., Browns Park at Colorado border, milemarker 97,2 June 1988, C. R. Nelson, #5214. 19 May 1988, T. Evans & B. A. Haws. No SCI taken. (1.8). 13. UTAH: Daggett Co., Browns Park 200m W Colorado 31. UTAH: Millard Co., 11.7 mi W Leamington, Hwy 6 border on N side Green River, 28 June 1988, C. R. milemarker 97,24 July 1989, B. A. Haws & C. R. Nelson, CRN #5258. (2.0). Nelson, CRN #5373. (1.8). 14. UTAH: Emery Co., 53 mi W Green River, I-70 mile­ 32. UTAH: Millard Co., 20 mi W Delta, H wy 6/50 at marker 111.5, 26 May 1988, C. R. Nelson, CRN #5209. power line, 2 June 1988, C. R. Nelson, CRN #5215. No SCI taken. (2.0). 15. UTAH: Emery Co., 53 mi W Green River, I-70 mile­ 33. UTAH: Millard Co., Snake Valley, 10 mi E Nevada marker 111.5, 13 July 1988, B. A. Haws & C. R. border, Hwy 6/50,2 June 1988, C. R. Nelson, CRN Nelson, CRN #5270. (1.6). #5216. (2.1). * 16. UTAH: Emery Co., San Rafael Swell, 3 mi E Castle 34. UTAH: Millard Co., Snake Valley, 6 mi E Nevada Dale, road to San Rafael River bridge, 26 May 1988, border, Hwy 6/50 milemarker 6, 24 July 1989, B. A. C. R. Nelson, CRN #5211. (3.0).* Haws & C. R. Nelson, CRN #5374. (2.8). 17. UTAH: Emery Co., San Rafael Swell, 3 mi E Castle 35. UTAH: Millard Co., Snake Valley, 4.5 mi E Nevada Dale, road to San Rafael River bridge, 14 July 1988, border, Hwy 6/50, Eskdale Junction, 2 June 1988, , B. A. Haws & C. R. Nelson, CRN #5273. (2.6). * C. R. Nelson, CRN #5217. (2.3). 18. UTAH: Emery Co., San Rafael Swell, 6 mi E Castle 36. UTAH: Millard Co., Snake Valley, 4 miN Gandy, Dale, road to San Rafael River bridge, 26 May 1988, Trout Creek to Callao road, 3 June 1988, C. R. C. R. Nelson, CRN #5212. (2.0). Nelson, CRN #5222. (1.8). 19. UTAH: Emery Co., San Rafael Swell, 6 mi E Castle 37. UTAH: Millard Co., 3 mi E Clear Lake, 3 mi W Dale, road to San Rafael River bridge, 14 July 1988, Pavant Butte, 14 June 1988, C. R. Nelson, CRN B. A. Haws & C. R. Nelson, CRN #5272. (1.8). #5241. (2.6). 20. UTAH: Grand Co., 1.5 mi N Crescent Junction, 38. UTAH: Millard Co., 5 mi W Clear Lake, 1.3 mi E Thompson Pass Road, 25 May 1988, C. R. Nelson & Hwy 257, 14 June 1988, C. R. Nelson, CRN #5242. T. Tibbetts, CRN #5205. No SCI taken. (1.8).

155 39. UTAH: Millard Co., 20 mi S Deseret, Hwy 257 mile­ 58. UTAH: Uintah Co., 5 mi S Bonanza, junction major marker 48, 14 June 1988, C. R. Nelson, CRN #5243. access road to Bookcliffs, on hillside, 29 June 1988, (1.2). C. R. Nelson, CRN #5261. (2.3). 40. UTAH: Millard Co., outside SE corner Desert Experi­ mental Range, inside first fence N ofHwy 21, 15 June RESULTS 1988, C. R. Nelson, CRN #5249. (2.5). Eight species of coccoid homopterans in four families 41. UTAH: Millard Co., outside E border Desert Experi­ were collected on shadscale during the survey (table 2). mental Range, 15 mi N Hwy 21, 15 June 1988, Of these four families, Ortheziidae and Pseudococcidae C. R. Nelson, CRN #5251. (1.5). were encountered most frequently during sampling. The 42. UTAH: Millard Co., 4 mi E Crystal Peak, 15 June 1988, other two families, Coccidae and Diaspididae, were rarely C. R. Nelson, CRN #5252. (1.8). encountered but were very abundant on the few sites where they were found. Another family of coccoid, Dacty­ 43. UTAH: Millard Co., 0.5 mi E Brown Knoll (near S end lopiidae, was frequently collected at the same sites but of Sevier Lake), 15 June 1988, C. R. Nelson, CRN only in association with prickly pear (Opuntia spp.). No #5253. (2.4). dactlyopiids were ever seen on shadscale during the sur­ 44. UTAH: Sevier Co., 1 mi E Fremont Junction, I-70 mile­ vey. It is interesting to note that no aphids (Homoptera: marker 90,26 May 1988, C. R. Nelson, CRN #5210. Aphididae and related families) were collected on shad­ (2.8). scale, both during the general survey of the insects of shadscale and the directed survey of coccoids. In the sys­ 45. UTAH: Sevier Co., 1 mi E Fremont Junction, I-70 mile­ tematic treatment that follows, the numbers in parenthe­ marker 90, 13 July 1988, B. A. Haws & C. R. Nelson, ses represent the locality numbers (given above) from CRN #5271. (1.9).* which the taxon was collected. 46. UTAH: Tooele Co., Rush Valley, 0.5 mi W Faust rail­ road station, 2 May 1988, C. R. Nelson, CRN #5166. No SCI taken.* Table 2-8cale insects and mealybugs 47. UTAH: Tooele Co., Rush Valley, 0.5 mi W Faust of shadscale in the Intermoun­ railroad station, 23 June 1988, C. R. Nelson & tain West (Insecta: Homoptera: R. Rasmussen, CRN #5256. (1.3). Coccoidea)

48. UTAH: Tooele Co., Rush Valley, 0.5 mi W Faust Coccidae railroad station, 5 August 1988, C. R. Nelson, CRN Ceroplastes irregularis Cockerell #5278 transect 1. (0.6). Diaspididae 49. UTAH: Tooele Co., Rush Vailey, 0.5 mi W Faust Aonidomyti/is incisus Ferris railroad station, 5 August 1988, C. R. Nelson, CRN Ortheziidae #5278 transect 2. (1.6). Orthezia annae Cockerell 50. UTAH: Tooele Co., Rush Valley, 0.5 mi W Faust Pseudococcidae railroad station, 16 August 1988, C. R. Nelson & Chorizococcus polyporus McKenzie D. L. Nelson, CRN #5283 transect 1. (1.8). Distich/icoccus salinus (Cockerell) Humococcus atriplicis Ferris 51. UTAH: Tooele Co., Rush Valley, 0.5 mi W Faust Phenacoccus solenopsis Tinsley railroad station, 16 August 1988, C. R. Nelson & Puto atriplicis McKenzie D. L. Nelson, CRN #5283 transect 2. (2.3). 52. UTAH: Tooele Co., Rush Valley, Faust, Pony Express Monument, 2 May 1988, C. R. Nelson, CRN #5167. No SCI taken. Coccidae 53. UTAH: Tooele Co., Rush Valley, Faust, Pony Express Monument, 23 June 1988, C. R. Nelson Ceroplastes irregularis Cockerell, 1893a:351 & R. Rasmussen, CRN #5257. (1.8). (fig. 3)-The wax scale family Coccidae is represented on shadscale by a single species, Ceroplastes irregularis 54. UTAH: Tooele Co., Rush Valley, 3.7 mi E Faust rail­ Cockerell. Published records of the hosts of this species road station, off-road site, 23 June 1988, C. R. Nelson include the chenopod saltbushes (Atriplex spp.), seepweed & R. Rasmussen, CRN #5255. (1.6). (Suaeda spp.), winterfat (Ceratoides lanata [Pursh] Moq. 55. UTAH: Tooele Co., Davis Mountain, 15 air mi W [as Eurotia]); the composites sagebrush (Artemisia spp.) Vernon, 2 May 1988, C. R. Nelson, CRN #5168. and rabbitbrush (Chrysothamnus spp.); and salt cedar No SCI taken.* (Tamarix sp.) (Essig 1931; Essig 1958; Gill 1988). During our surveys we found C. irregularis on shadscale and 56. UTAH: Tooele Co., Davis Mountain, 15 air mi W fourwing saltbush at sites in the Great Basin (4, on both Vernon, 3 June 1988, C. R. Nelson, CRN #5223. (2.9). shadscale and fourwing saltbush) and the Colorado Pla­ 57. UTAH: Uintah Co., 13 mi S Hwy 40, Old Hwy to teau (13, on shadscale; and at Utah: Grand Co., Dry Bonanza, 29 June 1988, C. R. Nelson, CRN #5260. Valley, on fourwing saltbush). (2.4).

156 Figure 3-Cerop/astes irregularis on shadscale. Idaho, Owyhee Co., locality 4. Scale bar= 1 em.

Figure 4-Aonidomytilis incisus on fourwing saltbush. Utah, Grand Co., Dry Valley. Scale bar= 1 em, ruler in millimeters.

At the Murphy, ID, site greasewood (Sarcobatus vermi­ that of Aonidimytilis, which is smaller, maximum length culatus [Hooker] Torrey) was abundant and mixed in with 2 mm; elongate, three times as long as wide; and silky, both shadscale and fourwing saltbush but was not in­ nonreflective white in color. fested with C. irregularis, which was extremely abundant Our infrequency of encountering this species leads us on the other two plant species. All are chenopod shrubs; to believe that it could not be a major cause of widespread the adult females of this species are sessile, attached to dieoff of shadscale although its presence may indicate the crown of the shadscale plant. The sessile forms en­ catastrophe for localized individual populations of the crusted most of the stems and branches at the Murphy shrub. site, under conditions of extremely high infestation. This species can be separated from all other coccoids occurring on shadscale, exceptAonidomytilis incisus, by its sessile Diaspididae growth form. It may be separated from Aonidimytilis Aonidomytilia inciaua Ferris, 1943:73 (fig. 4)--A by the irregularly globular form (diameter 3 mm and single species of the armored scale family Diaspididae was greater) of the wax secretion, which is generally brown found on shadscale, Aonidomytilis incisus. This species but occasionally glossy white. This form contrasts with

157 was previously known only from the original type series Ortheziidae collected on winterfat in Coconino County, AZ. We col­ lected this species from fourwing saltbush in Grand Co., Orthezia annae Cockerell, 1893b:403 (figs. 1, 5-7; UT, and on shadscale at localities 12 and 15. Other dia­ map: fig. B)-Ensign coccids were frequently encountered spidids thought to be conspecific, but unconfirmed by taxo­ during our surveys, often in large numbers (figs. 5-6). We nomic authorities, were collected from shadscale at local­ found the species Orthezia annae throughout both the ities 4 and 8 near the western margin of the Great Basin. Great Basin and the Colorado Plateau. In fact this spe­ The rare occurrence of this species eliminates it from con­ cies was more widespread than any other coccoid species sideration as a major factor in widespread shrub dieoff. encountered during this survey (fig. 8). Its widespread

Figure 5-0rthezia annae on shadscale. Utah, Beaver Co., locality 10. Scale bar= 1 em.

Figure 6-0rthezia annae on shadscale. Utah, Tooele Co., locality 47. Scale bar= 1 em.

158 ... '·~ Figure 7-0rthezia annae on shadscale. Utah, Millard Co., locality 39. Scale bar = 1 em.

egg mass (ovisac) attached to the posterior margin of their bodies (fig. 7). The large size (up to 1 em) of adult 0. annae relative to other mobile coccoids (Pseudococcidae) can also serve to separate the two groups. Many other taxonomic details of the family Ortheziidae are available (Morrison 1952). It should also be noted that 0. annae commonly feeds on fourwing saltbush, often in high population numbers. In contrast to the root-feeding habit of this species on shad­ scale, it feeds aboveground on fourwing saltbush. Particu­ larly severe infestations on fourwing saltbush were noted at sites along the Colorado River near Moab. The terminal .C) •• branches of the shrubs were attacked at these sites. Wide­ spread death of shrubs along with shrubs in various states .. ~ of decline was noted in association with these insects. These insects may indeed be a major cause of the decline and dying of shadscale and fourwing saltbush in the west­ ern United States.

Pseudococcidae

Figure a-Distribution of Orthezia annae, 1988-89 Five species ofPseudococcidae were encountered during sites. our survey of the root zone of shadscale. We found few characters that could be used to distinguish these species in the field. Attempts were made to correlate color, color of body fluids, type and distribution of waxy exudates, and distribution and high population numbers on selected overall shape with individual species within the Pseudo­ plants make this species a prime suspect in the dieoff prob­ coccidae for field identification. These attempts generally lem. The relationship of this species to the dieoffproblem failed, although we had limited success in field-identifying will be examined later in this report. Localities include: 1, the salt grass and the atriplex mealybug as 2, 7, 10, 13-15, 18-19, 39, 42, 44, 47-48, 50, 52-53, and 58. noted under their individual treatments. Two or more Orthezia annae actively, albeit slowly, crawl on the sur­ species of Pseudococcidae were regularly found on the tap­ face of the host plant. They can be readily separated from root of the same plant during our studies (fig. 9). other mobile coccoids on shadscale by the long, fluted waxy were often found tending the various pseudococcid plates which are exuded from and attached to their bodies. species. One small yellow species, Solenopsis molesta Their true shape (circular and flattened dorsoventrally) (Say), was very frequently encountered with several mealy­ is often masked by the waxy plates. Mature females are bug species and with the ortheziids. This species, in recognizable in having straight wax plates covering the the same genus as the pestiferous

159 Figure 9-Pseudococcidae on shadscale, mixed colony of Chorizococcus polyporus and Phenacoccus so/enopsis. Utah, Millard Co., locality 30. Scale bar= 1 em.

(Solenopsis invicta Buren) was rarely encountered in the encountered in a frequency and with adequate numbers absence of mealybugs during our sampling. The list of ant to consider it a possible agent in the widespread dieoff species we found tending Coccoidea on shadscale includes: of shadscale. In addition, this species was frequently Camponotus vicinus Mayr, Crematogaster mormonum collected from the same plants as Phenacoccus solenopsis, Emery, Formica sp., Lasius alienus Mayr, Myrmica amer­ another major suspect in dieoff. icana Weber, and Solenopsis molesta (Say). Ant identifi­ cations were made by C. R. Nelson using keys found in Wheeler and Wheeler (1986) and Allred (1982). Chorizococcus polyporus McKenzie, 1961:17 (fig. 9; map: fig. 10}--Chorizococcus is a large genus ofPseudo­ coccidae with at least 25 species in North America. The host range for the genus is broad, with records from such disparate plant families as Asteraceae, Asclepiadaceae, Cupressaceae, Fabaceae, Nyctaginaceae, and Poaceae (McKenzie 1967). Chorizococcus polyporus has been given the common name many-pored mealybug. It has been recorded from soil beneath juniper (Juniperus spp.) trees, which included roots of Eriogonum fasciculatum (Bentham) var. polifolium (Bentham) Torrey & Gray and Elymus elymoides (Rafinesque) Swezey (as Sitanion hystrix [Nuttall] J. G. Smith). Other host records (McKenzie 1967) .?J. include specimens from the roots of wild buckwheat (Eriog­ onum fasciculatum var. foliolosum [Nuttall] Stokes) and •• t baccharis (Baccharis sp.). This species has not previously been reported from salt­ bush and no members of the genus from California are recorded from Chenopodiaceae. The species is recorded only from California from two southern counties (Riverside and San Bernardino), which harbor many species of salt­ bush. We record this species from the roots of shadscale Figure 10-Distribution of Chorizococcus polyporus, (fig. 9) from numerous localities throughout Utah (10, 11, 1988-89 sites. 18, 20-21, 23, 30, 37-38, 40, and 56-58). This species was

160 Distichlicoccus salinus (Cockerell), 1902:21-Six Puto atriplicis McKenzie, 1961:33 (map: fig. 12)­ species of Distichlicoccus have been formally described The atriplex mealybug belongs to a large genus with broad (McKenzie 1967). We regularly encountered one species, host range. This species, however, is limited to the genus Distichlicoccus salinus-the salt grass mealybug-in our Atriplex and has been recorded from both California and samples. All known hosts of this genus are grasses (family, Idaho (McKenzie 1967). We found this species on the Poaceae), and the occurrence of D. salinus on shadscale roots of shadscale in a few localities near the eastern edge should be regarded as anomalous (R. J. Gill, personal com­ of the Great Basin in the general vicinity of Sevier Lake munication). Grasses, particularly bottlebrush squirrel tail Oocalities 28, 32, 35-36, 38, and 41). (Elymus elymoides [= Sitanion hystrix]) and Indian rice grass (Stipa hymenoides Roemer & Schultes[= Oryzopsis hymenoides (Roemer & Schultes) Ricker]), often were dug in close association with the crown of shadscale during this study. Extreme care was taken to separate the grasses from the shadscale before mealybugs were examined and counted. Despite this care, occasional individuals of D. salinus were taken from shadscale. We conclude that col­ lections of these individuals are probably spurious on shad­ scale, having either been dislodged from their proper host during sampling or merely traveling across the shadscale while changing position on their grass host. We could gen­ erally identify this species in the field by its small size, elongate shape, and bright orange coloration. Collection localities include: 21-22, 24, 40, and 57. Low frequency of encounter and small population size on shadscale (especially as it relates to improper host asso­ ciation) preclude this species from a major role in the dieoff problem. Humococcus atriplicis Ferris, 1953:371-Eight species of Humococcus are reported from North America (McKenzie 1967). Various species have been reported using saltbush, grasses, and composites as hosts. Humococcus atriplicis, the Ferris atriplex mealybug (not to be confused with the atriplex mealybug, Puto atriplicis), has been recorded from Figure 11-Distribution of Phenacoccus solenopsis, . California, Mexico, and Texas. In all cases its recorded host 1988-89 sites. is saltbush with a Death Valley record of it from desert holly saltbush (A. hymenelytra [Torrey] Watson). We col­ lected this species on shadscale from both the Great Basin and the Colorado Plateau (localities 19 and 37). The rare occurrence of the Ferris atriplex mealybug in our study area eliminates it as an agent in the widespread dieoff of shadscale. Phenacoccus solenopsis Tinsley, 1898:47 (fig. 9; map: fig. 11}-The solenopsis mealybug (McKenzie 1967) belongs to a large genus (more than 35 species in North America), which has an extreme host range. The host range of this particular species is also impressive, including saltbushes and other Atriplex, Asteraceae, Euphoriaceae, Malvaceae, Orobanchaceae, Solanaceae, as well as the chenopod genus Suaeda. Distributional records list it from Arizona, California, Colorado, Mississippi, New Mexico, Washington, DC, and Texas. We herein record it from nu­ merous localities in Utah (10, 20, 22, 24-25, 28, 30, 36-37, 39, 47, 53-54). The distribution of the solenopsis mealybug parallels that of the many-pored mealybug, and the two species often share the same individual host plant. This species has high potential for being an important agent in dieoff, since it was frequently encountered during sampling and regularly occurred in relatively high num­ bers. We did not detect this species from the Great Basin Figure 12-Distribution of Puto atriplicis, 1988-89 outside Utah; however we took only limited samples in sites. Idaho, Nevada, and Oregon.

161 The numbers of atriplex mealybug individuals on shad­ 12 scale roots were generally low. Limited success in identify­ ing adult females of this species in the field made use of 10 their large size, larger wax plates, and broad circulus. The infrequent encounter of the atriplex mealybug during our (/) 8 sampling eliminates it as a prime suspect associated with Q) -~ widespread dieoff of shadscale. CfJ 6 4 WIDESPREAD SHADSCALE DIEOFF AND ITS RELATIONSHIP TO 2 COCCO IDEA a~~~~~~~~~~~~~~~ ~N~~~~OOmN~~~~~~OOmM Shadscale Condition o~~~~~~~ NNNNNNNN One important ancillary of the survey of the coccoids of shadscale was a concurrent survey of the health of this SCI plant from widespread sites in the Great Basin and on Figure 13-Frequency of SCI values. the Colorado Plateau. We recorded a Site Condition Index (SCI) for many localities (CRN #5210-5380). Figure 13 •;..... summarizes the distribution of SCI values for the sites, the mean SCI value for the sites was 2.01 (SD = 0.53); four sites had SCI values of 2.0, while 24 had values less 40 than 2.0, and 20 had values greater than 2.0. We conclude, therefore, that the "average" stand of shadscale in the In­ Q) () termountain West would fit our category 2 condition (hav­ c 30 ing 34-66 percent living branches). Stated another way, Q) one-third ofthe branches of the shadscale plants sampled to... to... were dead, from unspecified causes. The potential increase ::J in production of this "dead third" may well be worth the () 20 () effort to ascertain the cause. 0 Prevalence of Coccoidea eft. 10 A diverse coccoid fauna has developed on shadscale. From amid the eight species of coccoids present on shad­ 0 scale some indication of incidence (percent plants infested) Orth Phen Char Puto Humo Dist Cere Aoni none and relative abundance is necessary to determine their role, if any, in the dieoff problem. Detailed taxonomic and Taxa numeric abundance data are available for 35 of our sites Figure 14-lncidence and relative abundance of (CRN #5210-5273). Additional data for Orthezia annae are species of Coccoidea on shadscale. available for the 13 remaining sites (CRN #5274-5384). The specific identity, however, of the pseudococcids and diaspidids occurring at these sites is not available. Using the data from the 35 sites, we can determine the incidence each of four comparisons: (1) SCI versus percent of plants and relative abundance for the coccoids (fig. 14). Three 2 infested with 0. annae at individual sites (r = 0.03, F 1 46 = of the coccoid species (Chorizococcus polyporus, 0. annae, 0.91, p < 0.65); (2) SCI versus number of 0. annae on all 2 Phenacoccus solenopsis) occur with enough frequency , plants at a given site (r = 0.14, F 1 46 = 7.60, p < 0.01); (3) (each, coincidentally occurring at 31 percent of the sites) SCI versus percent of plants infested with pseudococcids at to warrant further and more detailed investigation into 2 individual sites (r = 0.10, F 146 = 4. 74, p < 0.03); and (4) SCI their possible role in dieoff. The remaining five coccoid versus number ofpseudococcids on all plants at a given site species occur much more rarely and probably play lesser 2 (r = 0.06, F 146 = 2.91, p < 0.09). The relationship between roles in shadscale mortality, although the local role of the two variables in each of these comparisons was very Puto atriplicis in its endemic range should certainly be weak, with only a slight increase in numbers and incidence considered. It is also interesting to note that coccoids oc­ on the unhealthier plants. curred at 80 percent of the shadscale sites under consid­ eration, with only 20 percent of the sites being apparently Orthezia annae-Since this species is widespread and coccoid-free (fig. 14). Twenty-five of the 35 considered sites common throughout the range of shadscale, it may play an (71 percent) harbored one or more of the three important important role in the dieoff problem. Our sampling proce­ mealybug species. dures allowed us to determine if these insects were more We also performed separate analyses comparing SCI likely to occur on healthy or unhealthy plants. By combin­ with both the presence of 0. annae and that of all pseudo­ ing data from all sites we are able to make some generali­ coccids combined. We used a linear regression model for zations about the occurrence of 0. annae (expressed as

162 All Pseudococcidae 20

0 / 0 10

0 1 2 3 Condition Figure 17-lncidence of Pseudococcidae on shad­ scale, see text for condition category definitions.

few 0. annae (six total specimens on 348 juvenile plants 2000 sampled, average of 0.02 individuals per plant). The ca number of this species feeding on plants of each condition .N category increased with the health of the plants sampled. Q) Plants of condition 1 had 496 individual ensign coccids ..c. on 440 plants sampled (average of 1.13 individuals per t::: plant), condition 2 had 1,211 individuals on 443 plants 0 1000 sampled (average of 2. 73 individuals per plant), and con­ =1:1: dition 3 had 1,284 on 497 (average of 2.58 individuals per ca plant). Thus, juvenile plants harbored very few 0. annae, ~ condition 1 plants had a few more, and condition 2 and 3 0 plants harbored approximately three times the number of 1- 0. annae that the unhealthier plants of condition one did. In summary, using presence/absence data to correlate 2 3 percentage of plants infested with plants of the four condi­ tion categories we find that large plants are elected over Condition small ones and that healthy plants are elected over un­ healthy ones. These preferences are shown as well in the Figure 16-Number of Orthezia annae on shad­ scale, see text for condition category definitions. analysis summarizing the actual numbers of ensign coc­ cids found on the various plant conditions, with plants of condition 2 and 3 harboring 0. annae at about the same rate. percentage of plants harboring any specimens and as Pseudococcidae-We were unable to distinguish the total number of sp~cimens) on each condition category. species ofpseudococcids on shadscale in the field. We Figure 15 summarizes the incidence or percentage of plants analyzed the presence/absence of pseudococcids, the num­ of each condition harboring ensign coccids. Juvenile plants ber of individuals of all pseudococcid species combined, were used less frequently (0.085 percent infested) than and the average number of individual pseudococcids per plants of any other of the condition categories. Following plant in the same fashion as we did with the single spe­ juveniles, plants of increasingly better condition were more cies 0. annae. Results were similar. The percentage of likely to be found harboring ensign coccids. The healthiest plants infested with pseudococcids was low (fig. 17) on plants (condition 3) were most likely to have this species juvenile plants (4.3 percent), intermediate on plants in on their roots (10.25 percent) than plants of intermediate poor condition (condition 1, 11.8 percent), and high on the health (condition 2, 8.04 percent), followed by plants in the remaining, healthier plants (14.2 percent on condition 2 lowest plant health category (condition 1, 4.61 percent). and 14.3 percent on condition 3). The actual number of This trend is followed as well in the relationship between pseudococcids (fig. 18) sampled on shadscale was 229 indi­ the number of ensign coccids associated with each of the viduals on 348 juvenile plants (average of 0.66 individuals condition categories (fig. 16). Juveniles harbored very per plant), 658 on 440 plants of condition 1 (average 1.50

163 from the direct impact of the coccoids, or be sufficiently All Pseudococcidae weakened to die from additional factors, or would recover. 1500 In each case, however, the remains of coccoids will not persist (and allow detection) after their death, since they quickly decompose (C. R. Nelson personal observations) as a result ofliving in and near the soil with its large bac­ terial and fungal populations (Nelson and others 1990). ~ 1000 Using either of these two scenarios: of emigration or popu­ ctS lation collapse and subsequent specimen decomposition, +.J 0 coccoids would not have been on the unhealthy plants for +.J us to sample. These theoretical arguments emphasize the notion 500 that direct measurements of the impact of coccoids on shadscale need to be made. Controlled experiments are needed to determine what population levels of coccoids result in shrub mortality under a variety of environmen­ 0 tal conditions. Detailed biological studies are needed to 1 2 3 determine basic life-cycle data from which lifetables can Condition be constructed. Population dynamics experiments are . ·.··. needed to determine intrinsic growth rates and reproduc­ tive potential. During our surveys, several species of Figure 18-Number of Pseudococcidae on shad­ parasites and predators have been associated with the scale, see text for condition category definitions. coccoids. Studies of these natural control agents should be made to determine how their populations might be augmented. individuals per plant), 1,207 individuals on 443 plants The interesting contrast between belowground feeding of condition 2 (average of2.72 individuals per plant), and of 0. annae and Ceroplastes irregularis on shadscale and 1,223 individuals on 497 plants of condition 3 (average aboveground feeding by these species on fourwing salt­ of 2.46 individuals per plants). Juvenile plants harbored bush should be further investigated. Knowledge of envi­ few pseudococcids, although they harbored the combined ronmental factors regulating this differential architecture pseudococcid species at a rate considerably higher than preference on related host plants may allow for prediction they did the single species 0. annae. Plants of condition of outbreaks using climatological and remotely sensed 1 harbored substantially more pseudococcids than did the data. juveniles but not as many as did either condition 2 or 3 Infestation rates of coccoids on shadscale of different plants, which bore pseudococcids at about the same rate. genetic makeup (in particular, ploidy) should be inves­ tigated. Ifmonoculture stands of uniform ploidy are at greater risk of decline and mortality from coccoid popu­ Conclusions of Prevalence Analysis lations, then control measures could be implemented. More 0. annae and pseudococcids are found on large, Our results show that several species of coccoids do healthy plants than on small, unhealthy ones. From this indeed occur throughout the range experiencing dieoff. it might be concluded that these insects are not responsible This, and the numbers of individuals found on many of for the mortality of plants seen during the present dieoff the plants sampled, leads us to conclude that coccoids are event, since high populations of the insects are not associ­ playing a major role in the mortality of shadscale. ated with plants of poor health. We conclude differently using the following reasoning. As demonstrated above, CONCLUSIONS coccoid insects thrive on the healthiest plants, perhaps since the healthiest plants can supply the most nutrients. Three of the eight species of coccoid Homoptera col­ Theoretically, through the feeding of the coccoids the lected on shadscale (Chorizococcus polyporus, Orthezia condition of the healthy plant would degenerate to a point annae, and Phenacoccus solenopsis) are widespread and where it could no longer support the coccoids. At this point abundant throughout the sampling area of the Great the coccoids could either emigrate from this host plant to a Basin and the Colorado Plateau. A fourth species, Puto more suitable, healthy plant or remain with the first plant atriplicis, occurs in reasonably high numbers but in a and die. If the coccoids emigrate (leaving the plant with no more limited area. Eighty percent of the sites surveyed insects for us to detect) the plant might be able to recover had plants infested, to varying degrees, with coccoids. from the infestation if other existing conditions permit, or Poor shrub health is only weakly correlated with either the plant would die without any trace of coccoids (leaving high incidence of coccoids or high numbers of coccoids. no specimens which we could detect). Further research Small, juvenile plants are less likely to harbor coccoids into the biology of the coccoids on shadscale is necessary than are larger, mature plants. Healthier plants are to determine migration rates. If the coccoids were to re­ more likely to host coccoids than are unhealthy plants. main with the now unsuitable plant they would eventually Healthy plants harbor generally larger populations of die, while at the same time further reducing the viability these insects. of the host plant. The host plant would then either die

164 Additional surveys are desirable, particularly from far­ Borror, Donald J.; Triplehorn, Charles A.; Johnson, ther south in the range of shadscale, to determine how Norman F. 1989. An introduction to the study of in­ closely the range of the various coccoids coincides with sects. 6th ed. Philadelphia: Saunders College Publish­ their host and how increasing temperatures might affect ing. 875 p. them. Biological and environmental factors regulating Cockerell, T. D. A 1893a. Three new Coccidae from coccoid populations should be investigated in greater the arid region of North America. Entomologist. 26: detail. 350-352. Cockerell, T. D. A. 1893b. Two new Coccidae from New Mexico. Annals and Magazine of Natural History. ACKNOWLEDGMENTS (6)12: 403-406. The provision of financial, human, and material re­ Cockerell, T. D. A 1902. New genera and species ofCoc­ sources of the Bureau of Land Management has added cidae, with notes on known species. Annals and Maga­ greatly to the scope of this project. This agency is grateful­ zine of Natural History. (7)9: 20-26. ly acknowledged. This research was accomplished through Essig, E. 0. 1931. A history of entomology. New York: a cooperative agreement between Utah State University Macmillan. 1029 p. and the USDA Forest Service, Intermountain Research Essig, E. 0. 1958. Insects and mites of western North Station (INT-87281). We especially thank Dr. Raymond J. America. New York: Macmillan. 1050 p. Gill of the California Department of Food and Agriculture Ferris, G. F. 1943. Additions to the knowledge of the Dia­ for identification of the coccoids that are herein reported. spididae (Homoptera:Coccoidea). Microentomology. 8(2): We also thank him for suggestions regarding host specifi­ 58-79. cities and literature citations, as well as providing unpub­ Ferris, G. F. 1953. Atlas of the scale insects of North lished host records and his review of the manuscript. We America. Vol. VI: 371-372. Palo Alto, CA: Stanford also thank Dr. Stanley L. Welsh of the M. L. Bean Life University Press. 278 p. Science Museum at Brigham Young University for pro­ Gill, Raymond J. 1988. The scale insects of California, viding logistic support to the senior author during the part 1: the soft scales (Homoptera; Coccoidea; Coccidae). early stages of this research. Thanks are also given to California Department of Food and Agriculture Techni­ the Tilton Fellowship and Department of Entomology at cal Series in Agricultural Biosystematics and Plant Pa­ the California Academy of Sciences, San Francisco, for thology No. 1. 132 p. enabling the senior author to continue this research as Little, V. A. 1972. General and applied entomology. 3d ed. well as to attend the symposium where these results were New York: Harper & Row. 527 p. initially presented. The Department of Zoology, University McKenzie, Howard L. 1961. Second taxonomic study of of Texas at Austin, is gratefully acknowledged for support California mealybugs, with descriptions of new species through the Brackenridge Field Laboratory. We thank (Homoptera: Coccoidea: Pseudococcidae). Hilgardia. 31: Richard J. Patrock, Department of Zoology, University of 15-52. Texas, for helpful review comments and interpretive ad­ McKenzie, Howard L. 1967. The mealybugs of California. vice. We also thank Ted Tibbetts of Moab, UT; Richard W. Berkeley: University of California Press. 525 p. Baumann of Brigham Young University; Ryan Rasmussen Morrison, H. 1952. Classification of the Ortheziidae, sup­ of Provo, UT; and Kaye Nelson of Austin, TX, for help in plement to classification of the subfamily Ortheziinae. gathering part of the data upon which this study is based. Tech. Bull. 1052. Washington, DC: U.S. Department of Agriculture: 1-80. Stutz, H. C.; Sanderson, S.C. 1990. [Unpublished data]. REFERENCES Provo, UT: Brigham Young University. Allred, Dorald M. 1982. Ants of Utah. Great Basin Tinsley, J.D. 1898. An ants'-nest coccid from New Mexico. Naturalist. 42: 415-511. Canadian Entomologist. 30: 4 7-48. Borror, Donald J.; DeLong, Dwight M.; Triplehorn, Wheeler, George; Wheeler, Jeanette. 1986. Ants of Charles A. 1976. An introduction to the study of Nevada. Los Angeles, CA: Natural History Museum, insects. 4th ed. New York: Holt Rinehart and Winston. Los Angeles County. 138 p. 852p.

165