Aliso: A Journal of Systematic and Floristic Botany Aliso: A Journal of Systematic and Floristic Botany Laboulbeniales on Semiaquatic Hemiptera. III. Rhizopodomyces Laboulbeniales on Semiaquatic Hemiptera. III. Rhizopodomyces

.-Thaxter's original description of Rhizopodomyces is emended to reflect new information gained from a study of not only the type species, R . merragatae, but also of six additional species described as new, R. basifurcatus, R. californicus, R. erectus, R. geniculatus, R . mexicanus, and R. polhemusii. All of the species have been found on semiaquatic bugs of the family Hebridae (Hemiptera). The female thallus of Rhizopodomyces spp. consists of a simple, two-celled receptacle, the suprabasal cell bearing a one-celled primary appendage and giving rise usually to a single, long stalked perithecium. The male thallus initially consists of three cells, a two celled receptacle and a terminal cell which may form a simple antheridium directly or remain sterile and form a simple appendage, in which case the suprabasal cell of the receptacle divides and forms one or two simple an theridia directly. Rhizopodomyces is compared with several other dioecious genera of Laboulbeniales of similar morphology, Amorphomyces, Dioico myces, Dicrandromyces, Triandromyces, and Tetrandromyces. Specula tions regarding the possible relationship of Rhizopodomyces to these genera are presented.


Richard K. Benjamin
Abstract. -Thaxter's original description of Rhizopodomyces is emended to reflect new information gained from a study of not only the type species, R . merragatae, but also of six additional species described as new, R. basifurcatus , R. californicus, R. erectus , R. geniculatus , R . mexicanus , and R . polhemusii. All of the species have been found on semiaquatic bugs of the family Hebridae (Hemiptera). The female thallus of Rhizopodomyces spp. consists of a simple, two-celled receptacle, the suprabasal cell bearing a one-celled primary appendage and giving rise usually to a single, longstalked perithecium. The male thallus initially consists of three cells , a twocelled receptacle and a terminal cell which may form a simple antheridium directly or remain sterile and form a simple appendage, in which case the suprabasal cell of the receptacle divides and forms one or two simple antheridia directly . Rhizopodomyces is compared with several other dioecious genera of Laboulbeniales of similar morphology , Amorphomyces, Dioicomyces, Dicrandromyces, Triandromyces, and Tetrandromyces. Speculations regarding the possible relationship of Rhizopodomyces to these genera are presented.
In the first of my earlier papers on Laboulbeniales parasitizing semiaquatic Hemiptera (Benjamin 1967(Benjamin , 1970, my treatment of the classification of Hemiptera followed that outlined by Brues, Melander, and Carpenter (1954) in which the order is divided into two suborders, Homoptera and Heteroptera. I am now following a system preferred by many entomologists which recognizes two orders, Homoptera and Hemiptera (=Heteroptera) (Comstock 1947 ;Borror, DeLong, and Tripelhorn 1976). In the latter treatment, the Hemiptera are divided into three suborders: (1) Hydrocorizae, longhorned bugs (Cryptocerata of authors) which are mostly aquatic or semiaquatic; (2) Amphibicorizae , short-horned bugs (Gymnocerata, in part, of authors) which are primarily semiaquatic or shore inhabiting; and (3) Geocorizae, short-horned bugs (Gymnocerata, in part, of authors) which are usually terrestrial. With regard to Laboulbeniales, no representative has as yet been found on any of the Homoptera, whereas several have been described on Hemiptera.
Coreomyces Thaxter (1902), with 19 described species, is restricted to Corixidae (water boatmen), and members of this genus are the only Laboulbeniales known on Hydrocorizae, which includes also the backswim-ing as a sterile primary appendage, the suprabasal cell dividing and forming one or two simple, flask-shaped antheridia and a subtending sterile cell.
Female : Consisting initially of three superposed cells; the terminal cell forming a simple primary appendage; the basal and suprabasal cells constituting the receptacle; the suprabasal cell giving rise distally to a divergent branch bearing a terminal perithecium; the basal cell, above the point of attachment to the host, often forming laterally divergent, lobate or rhizoidlike, nonseptate outgrowths. Perithecium with well-developed stalk cells and basal cells; basal cells persistent; outer wall consisting of three vertical rows of four cells each and one vertical row of five cells. Spores once septate.
Type species: Rhiwpodomyces merragatae Thaxt. The thalli of species of Rhiwpodomyces typically develop on the upper surface of the host, mostly on the pronotum or hemelytra, although thalli may occasionally be found on the head, lower surface, or legs. The foot, i.e., the modified part of the basal cell (I) by which the fungus is attached to the host, is more or less appressed to the insect cuticle and may or may not become somewhat blackened ( Fig. 1D; 2J; 4A, C-J; 5A-C, G-M; 6B, C, D, I, H, L-O). Host penetration appears to be effected by a simple haustorium which is easily broken at the juncture of fungus and host when a parasite is removed. In whole mounts, haustoria could not be detected in the more or less transparent hemelytra at the point _ of attachment of a fungus although a small, circular region centrally located at the base of the foot marks the point of entry of the haustorium into the host cuticle ( Fig. 2D-

1, K).
Ascospores, where these could be observed in situ on the host, appeared always to have been discharged in pairs as is usual in most monoecious and dioecious Laboulbeniales . Mature ascospores observed inside perithecia (Fig. IF,I;2L;4B,M,N;5F,N) did not exhibit marked dimorphism as has been observed in some dioecious species (Benjamin 1970: Fig. 3D). The spore destined to become the female individual enlarges more rapidly as it develops than does the one forming the male ( Fig. 2A-D). Paired males and females remain closely associated throughout their development, and the two sexes often adhere to one another at the base when removed for study ( Maturation begins when the thalli reach the three-cell stage, which involves simply a single transverse division of the lower spore segment into the basal cell (I) and suprabasal cell (II) of the receptacle ( Fig. 2E ; 6B) . In the male of five of the species studied , the terminal cell, i.e. , the upper spore segment, is transformed into a single flask-shaped antheridium (   (m , n , n ') to the four rows of outer wall cells. Note protrudent fifth cell of the fivecelled row . Other details of perithecial cellular structure omitted. x 895 .-C. Male individual adherent to female at base near point of attachment of fungus to host. x 660.-D-F. Costa Rican specimen (RKB 2698).-D . Mature female and associated male (broken distally) . x 400.-E. Perithecium in lateral view showing terminus of primary stalk cell (VI) , secondary stalk cell (VII), basal cell m from which has arisen the posterior row of four wall cells, and basal cell n from which have arisen one lateral row of four wall cells and the anterior row of five wall cells (w'-w 5 ). Cell five of the anterior row (w 5 ) protrudes well beyond the apices of the terminal divides transversely and forms one or occasionally two distal antheridia having divergent efferent tubes (Fig. 4D,F;6C,D,H,I,M). The primary appendage of the female has been observed in seven of the eight species studied, and it is always simple, more or less elongate, and somewhat clavate . It may be missing (Fig . IA, D; 2J; 4A, G; 5A, C; 6D, H), be more or less persistent ( Fig. 2K; 5G, L) although sometimes shriveled ( Fig. IG; 4J), or the distal part may slough away and only the basal part remain (Fig. 4C, E; 61, 0).
Material adequate for a study of the salient features of perithecial development was available only for R. californicus. The perithecial primordium arises from the upper end of cell II immediately below the primary appendage ( Fig. 2F) which is shifted laterally and diverges from the upper end of the receptacle during maturation of the perithecium (Fig. 2F-H, K). The perithecial branch divides distally ( Fig. 2F; 3A), forming a small terminal cell (i), the primordial cell of the procarp from which will be derived the female sexual organ, and an elongate cell (h), the primordial cell of the perithecium from which will arise the stalk cells and basal cells of the perithecium. Cell i divides transversely and forms the trichophoric cell (tc) above and the carpogenic cell (cp) below ( Fig. 2G; 3B). The trichophoric ,cell elongates distally and cuts off a slender, more or less recurved, simple trichogyne (tr) ( Fig. 2H; 3C, D). Subsequently, the inner and outer wall cells of the perithecium will grow upward and around· the carpogenic cell ( Fig. 3G-I) which constitutes the carpogonium and from which the asci will derive (Fig. 3J).
Development of the perithecial wall begins with the delimitation of a small cell (j) which is cut off diagonally from the upper end of cell h ( Fig. 2G ; 3B). Cell h then cuts off a second small cell (m) which lies more or less adjacent to cell j ( Fig. 2H; 3D). Cell h now is termed the primary stalk cell of the perithecium (VI). Cell j then delimits two smallish cells (n, n') which are displaced upwardly to the right and left so that they lie more or less side by side (Fig. 21-K; 3E). Cell j now is termed the secondary stalk cell of the perithecium (VII). Cells m , n, and n' constitute the basal cells of the perithecium .
Perithecial wall development continues with the delimitation by cells m , n, and n' of the primary wall cells (o) (Fig. 3E). In R. californicus these cells elongate considerably during which time the trichogyne shrivels and is carried downward on the posterior side of the perithecium, the carpogenic cell (cp) elongates and the first parietal or inner wall cells (pc), which also arise from the basal cells, are formed between the carpogenic cell and the primary wall cells (Fig. 3F-G). No attempt was made in this study to trace the further development of the inner wall cells or the ascogenous system. As the perithecium elongates, additional outer wall cells (w 1 -w 5 ) are delimited ( Fig. 3H-J) until finally there are three vertical rows of four cells and one of five cells (Fig. 3K-M). Cell m gives rise to a single posterior row of four cells which always lies adjacent to the remnant of the trichogyne borne on the surface of the perithecium ( Fig. 2K; 31, K); cell n gives rise to two rows of cells, a lateral row of four and an anterior row of five ( Fig. 3K-M) ; and cell n' gives rise to the other lateral row of four cells (Fig. 3K-M). The position of cells n and n' relative to cell m may vary, n being dextral and n ' sinistral in one individual and vice versa in another individual.
In maturing females of all the species of Rhizopodomyces studied, the lower cell of the receptacle produces two or three basal protrusions that may elongate and become more or less rhizoidlike (Fig. IA , D, G; 2J, K; 4A , C, E , G, H, J; 5A , C, G , I, L ; 6H , I, L, N, 0). These outwardly extending structures appear to maintain close contact with the host surface but nowhere penetrate the integument. They most likely serve as buffer organs or props that help maintain the thallus in an erect posture.

Key to the species of Rhizopodomyces
A. Basal cell m of perithecium growing upward distally and forming a conspicuous free appendage 8. Rhizopodomyces sp. Basal cell m not forming a free appendage B B. Primary stalk cell of perithecium one half to two or more times longer than perithecium and its included basal cells C Primary stalk cell of perithecium usually shorter than perithecium and its basal cells E C. Stalk cell usually 2-3 times longer than perithecium; cell 5 of anterior row of outer wall cells of perithecium forming a fingerlike projection at perithecial apex l. R . merragatae Stalk cell less than 2 times longer than perithecium ; cell 5 of anterior row of wall cells not conspicuously protrudent D D. Perithecium geniculate; basal cell m of perithecium conspicuously enlarged and rounded externally; receptacular rhizoids poorly developed 3. R. geniculatus Perithecium not geniculate, cell m not enlarged ; receptacular rhizoids two, well developed, projecting at an acute angle from one another 7. R. basifurcatus E. Perithecium not geniculate, its long axis continuous with that of its primary stalk cell, its posterior and anterior margins about equally convex 2. R. californicus Perithecium more or less geniculate, posterior margin more strongly convex than anterior margin which may be nearly straight or slightly concave F F . Basal cells of perithecium prominent, relatively large and externally rounded, especially cells m and n 5. R . mexicanus Basal cells of perithecium not conspicuously rounded externally G G . Receptacle and primary stalk cell of perithecium forming a nearly straight axis; basal cell of receptacle two times longer than wide 4. R. erectus Receptacle and primary stalk cell of perithecium more or less geniculate ; basal cell of receptacle no longer than wide 6. R. polhemusii l. Rhizapodomyces merragatae Thaxt., Mem. Amer. Acad. Arts Sci. 16: 60, 1931, emend. FIG. 1 Male: Hyaline , faintly yellowish; receptacle nearly uniform in diam, attenuated near base; basal cell curved, about twice as long as the suprabasal cell ; antheridium terminal , only slightly shorter than the basal cell. Total length 38-43 µm; ca. 4-5 µ,m wide.
Female: Hyaline, faintly yellowish, usually somewhat curved, especially below ; basal cell only slightly longer than wide, 15-25 x 10-18 µ,m , with 2-3 basal rhizoidlike prolongations radiating irregularly, these 30-50 µ,m long, 9-12 µ,m wide at the base, 5-8 µ,m wide near the tip; suprabasal cell up to three or more times longer than the basal cell , 35-95 x 10-25 µ,m, bearing distally the laterally divergent, simple primary appendage; well-developed appendages not observed; primary stalk cell of the perithecium elongate, two times or more as long as the suprabasal cell, with which its long axis is continuous, 107-165 x 10-18 µ,m ; perithecium in line with the long axis of the primary stalk cell, its basal cells well defined and usually somewhat rounded externally, its posterior and anterior margins about equally convex , tapered distally, 48-77 x 16-30 µ,m , including basal cells; Thaxter's description of this species was based on two mature females and their accompanying males. Both specimens were figured by Thaxter (1931, Pl. XIV, Figs. 17-18) and, according to records accompanying the Thaxter Collection at the Farlow Herbarium, the individual shown in his The type collection of R . merragatae was found on Merragata brevis, whereas the other two collections cited above were found on species of Hebrus . These two genera of small, shore-inhabiting bugs are differentiated primarily on antenna] characteristics. They are otherwise similar and their separation may be for no other good reason than taxonomic convenience (Usinger 1956). Of two mature specimens from the Costa Rican collection, only one was intact and is shown in Fig. 1D-F (RKB 2698). Except for having two rather than three basal rhizoidlike prolongations and straight rather than slightly spiraled rows of outer wall cells this specimen seems hardly separable from the type. Three mature females make up the Mexican collection (RKB 2705A), and of these only one ( Fig. lG-1) is in good condition. Except for being somewhat larger, these specimens seem sufficiently like the type to be identified with it.

Rhizopodomyces californicus
Female: Hyaline, whitish; usually somewhat curved, basal cell about two times longer than wide, 12-30 x 8-16 µ,m , with two to three basal lobate or more or less elongate projections radiating irregularly, these 6-22 µ,m long, ca. 5-8 µ,m wide; suprabasal cell usually two times as long as the basal cell, 30-50 x 9-16 µ,m, bearing distally the laterally divergent, simple, straight or slightly curved, clavate primary appendage, 20-30 x 5-10 µ,m; primary stalk cell of the perithecium usually only slightly longer than the suprabasal cell, with which its long axis is continuous, 35-60 x 9-13(-16) µ,m; perithecium in line with the long axis of the primary stalk cell , its basal cells well defined and barely rounded externally , its posterior and anterior margins about equally convex, tapered distally , 45-60 ( Rhizopodomyces calif ornicus is the only species discussed in this paper for which an adequate sample of specimens was available for determining the probable true range of variation of thallus dimensions (31 females from the type locality were measured) and for observing early stages of development as discussed above. This species most closely resembles R. m erragata e in the conformation of its perithecium. In both species the posterior and anterior perithecial margins are about equally convex and cell five of The primary perithecial stalk cell of R. merragatae is two or more times as long as either its perithecium or the suprabasal cell of its receptacle, whereas in R. californicus the stalk cell and suprabasal cell are about equal in length. In the latter species the perithecium commonly is longer than its primary stalk cell (Fig. 2J, K). On the basis of material at hand for study, the prolongations arising from the basal cell of R . merragatae (Fig. IA, D, G) are more elongate and rhizoidlike than those of R. californicus ( Fig. 2I-K) . Benjamin, sp. nov. FIG. 4A-F Mas: Hyalinus, albidus , 31-36 µ,m in tota longitudine ; diametrum thalli prope uniforme , cellulis leniter extra rotundatis; cellula terminalis sterilis appendicem simplicem elongatam plus minusve clavatam ca. 8-9 x 4.5 µ,m formans; cellula suprabasilaris receptaculi 1-2 antheridia et cellulam suprabasilarem secundariam formans; cellula basilaris receptaculi 13-15 x 5.5-6.5 µ,m.
Male: Hyaline, whitish, nearly uniform in diameter, cells slightly rounded externally between septa; terminal cell sterile, forming a simple, elongate, more or less clavate appendage ca. 8-9 x 4.5 µ,m ; suprabasal cell dividing distally and forming one or two superposed antheridia having divergent efferent tubes; the resulting suprabasal cell and the basal cell subequal. Total length to tip of antheridium 31-36 µ,m; ca. 5.5-6.5 µ,m wide.
Female: Hyaline, whitish , uniformly curved from foot to base of perithecium; basal cell about two times longer then wide, 20-30 x 12-18 µ,m, with one or two basal projections, these hardly longer than wide, 8-10 x 6-7 µ,m; suprabasal cell about two times longer than the basal cell, 40-55 x 10-13 µ,m, with a more or less conspicuous subapical, posterior protrusion subtending the primary appendage; intact appendages not observed; primary stalk cell of the perithecium about 1.5 times longer than the suprabasal cell, with which its long axis is continuous, 65-75 x 9-14 µ,m; perithecium geniculate, its posterior margin strongly convex, its anterior margin concave, its basal cells well defined, cell m strongly convex externally and ca. 2-2.5 times longer than the secondary stalk cell , tapered distally, the apical ½ to ½ bent downward slightly, 44-57 Three mature females with their accompanying males from the type locality ( Fig. 4A, C, E) and one mature female from Arizona (Fig. 4G) constitute the material upon which the description of R . geniculatus is based. The strongly geniculate perithecium with its prominent, externally rounded basal cell m (Fig. 4B) readily distinguishes the species from its congeners. Each of the three males observed bore a somewhat shriveled appendage Male: Hyaline, pale yellowish, basal cell attenuated below , somewhat longer than the suprabasal cell, which is wider and slightly rounded externally; antheridium terminal, only slightly longer than the suprabasal cell. Total length 39 µ.m; ca. 6-7 µ,m wide.

Rhizopodomyces mexicanus
Male: Hyaline, receptacle nearly uniform in diam ; slightly constricted at the septa; basal cell about two times longer than the suprabasal cell, antheridium terminal, about half again as long as the suprabasal cell. Total length 44-46 µ,m; ca. 5.5-6 µ,m wide.
Female: Hyaline, pale yellowish, axis formed by the receptacle and primary stalk cell curved slightly; basal cell about two times longer than wide, slightly inflated above, 22-25 x 13 µm, with two basal rhizoidlike lateral prolongations 50-72 µ,m long, 7-8 µ,m wide at the base, 4 µ,m wide near the tip ; suprabasal cell about one-third longer than the basal cell, 30-32 x 10-13 µ,m; primary appendage not observed; primary stalk cell of the perithecium about as long as the suprabasal cell, 27-32 x 10-12 µ,m; perithecium moderately geniculate, its basal cells well defined and prominently rounded externally, especially cell m , its posterior margin strongly convex , its anterior margin slightly concave, tapered to its unmodified apex Five females , four mature and one immature, and three males (RKB 2705B) of this species were found on the same host that bore specimens here identified as R. merragatae (RKB 2705A; Fig. lG-1). Two of the mature females were in poor condition, one having the basal part missing. The male of R . mexicanus is relatively very large , nearly equaling the length of the receptacle of the female (Fig. 5A) . Although the two females illustrated (Fig. 5A, C) bear only one intact or broken rhizoidal prolongation, they have scars on the basal cells marking the location of second rhizoids broken off, probably, when the specimens were being removed from the host for mounting. The immature female, not figured, clearly has two basal rhizoids. The perithecium with its strongly convex posterior margin, slightly concave anterior margin, and especially the prominent basal cells (Fig. 5D-E) distinguish R . mexicanus from the other species. Unfortunately , the primary appendage was absent from all females. Benjamin, Hyalinus, 23-30 µ,m in tota longitudine; cellulae receptaculi subaequales externe rotundatae, cellula basilari 7-9 x 7-8 JLm, cellula suprabasilari 6-9 x 6-7 JLm; antheridium terminate 10-13 x 5-8 JLm.
Female : Hyaline, pale yellowish, the posterior and lateral median surfaces of the perithecium with a brownish-orange suffusion ; axis formed by the receptacle and primary stalk cell more or less geniculate; basal cell no longer than wide, 7-10 x IO JLm, with two nearly opposite, basal, rhizoidlike prolongations 44-55 JLm long, ca. 7 JLm wide at the base , ca. 5 JLm wide near the tip; suprabasal cell slightly wider than and two times longer than the basal cell, 20-22 x 13-17 JLm ; primary appendage broadly clavate to obovate, 18 x 12-13 JLm, borne near the middle of the posterior side of the suprabasal cell from which it projects nearly at right angles; primary stalk cell of the perithecium 1.5 to 2 times longer than the suprabasal cell , 33-40 x 12-17 JLm ; perithecium moderately geniculate , its basal cells well defined but inconspicuous, its posterior margin strongly convex, its anterior margin nearly straight or very slightly rounded below , tapered to the rela- Three females and their accompanying males (Fig. 5G-M) constitute the material upon which this species is based . Rhizapodomyces polhemusii is somewhat smaller than the other species and is the only one displaying any marked pigmentation other than around the foot. The brownish-orange suffusion involving the posterior and lateral surfaces of the perithecium is uni-formly present on all three females. The primary appendage is present on two of the females and is unlike that observed on other species in being broadly clavate to obovate and in projecting almost at right angles from near the middle of the suprabasal cell (Fig. 5G, L). I was unable to determine the exact number and arrangement of the outer wall cells, but it is doubtful that they deviate from the pattern established in the other species. Male : Mature individuals not seen. Female : Pale yellowish , uniformly curved from foot to tip of perithecium; basal cell slightly shorter than wide , 18-20 x 22-27 µ,m , with two broadbased rhizoidlike prolongations diverging laterally on the anterior side at an acute angle from one another, these 45-55 µ,m long , 12-15 µ,m wide at the base , 4-5 µ,m wide near the tip; suprabasal cell ca. three times longer than the basal cell , 60-70 x 18-25 µ,m , abruptly narrowed distally at the point of attachment of the nearly erect primary appendage; appendage nearly cylindrical with a rounded apex , 35-37 x 9-10 µ,m ; primary stalk cell of the perithecium about two times longer than the suprabasal cell, with which its long axis is continuous, 100-130 x 22-25 µ,m ; perithecium with well-defined basal cells which are somewhat rounded externally , its posterior margin more strongly convex than its anterior margin, uniformly tapered to its apex , 63-75 x 30--40 µ,m; outer rows of wall cells nearly straight; posterior and lateral rows consisting of four cells each ; anterior row consisting of five cells , the terminal cell projecting slightly beyond the apices of the terminal cells of the other rows.  (Fig. 60); slide in RSA.
The material upon which this species is based is not in the best condition. It consists of three mature and three immature female s and several paired immature males and females in various stages of development (Fig. 6K-N). Nevertheless, the fungus clearly is distinct from the other species and I have given it a name although I suspect that when more material becomes available for study its description, like that perhaps of several other taxa included in this paper, will need some emendation. The receptacle of R. basifurcatus is distinguished especially by its very short and wide basal cell bearing two broad-based, attenuated prolongations that diverge from one another at a rather close angle (Fig. 6L, N, 0). In one immature pair, the suprabasal cell of the male is growing upward immediately below the terminal cell (Fig.  6M), and it is likely that in this species, as in R. geniculatus , the terminal cell of the original spore forms a sterile appendage, the antheridium being derived from the suprabasal cell. This cannot be verified until mature males are discovered.
One of the females bore two perithecia, the stalk cell of the second having In the male, the upper of the two cells derived from the suprabasal cell (see B) has been converted into an antheridium (an) having an upwardly directed efferent tube; the still-persistent appendage (app) has been diverted on one side. x 895.-D. Immature female and accompanying male. The perithecium is in an early stage of development, but cell m is beginning to elongate. x 660.-E-G. Three sequential stages of development of the perithecium. In E, cell m has not yet cut off its primary wall cell. In F the upward development of m has begun, and in Git is beginning to elongate. E, x 1,080; F, G, x 1,360.-H. Immature female and accompanying male. The perithecium still is in an early stage of development , but the upward growth of cell m is well advanced. x 660.-I. Immature female and accompanying male. The outer walls of the perithecium have reached the four-cell stage, but the perithecium still is immature. The appendage derived from cell m probably has reached maximum size . Its base is adnate to the lower surface of the perithecium on the side away from the viewer. arisen immediately below that of the first. Whether or not the secondary perithecium had arisen following abortion or damage to the primary could not be determined. This is a phenomenon often observed in Laboulbeniales. In another collection of Rhizopodomyces in my collection (RKB 2701) specimens of female individuals that could not be positively identified (? nr. R. merragatae) bore the remnants of the stalk cells of several successive perithecia that had arisen from the suprabasal cell. The production of two or more successive perithecia in some species of Rhizopodomyces may prove to be of regular occurrence but additional observations are needed to resolve this question. There were no mature females, so the species cannot be given a name at this time. A unique feature of this taxon is the upward growth of basal cell m soon after perithecial development has begun (Fig. 6D-H). By the time the perithecium has reached the stage where four tiers of outer wall cells have formed, cell m has developed an elongate, nonseptate appendage (85-100 x 11-16 µ,m in the immature specimens observed) which is longer than and nearly as wide as the receptacle (Fig. 61-J).
Like R. geniculatus and perhaps R. basifurcatus, the terminal spore segment of the male of Rhizopodomyces sp. forms a sterile appendage, the antheridium developing from the upper of two cells derived from the suprabasal cell of the receptacle (Fig. 6B , C). A shriveled but persistent appendage was still attached to males accompanying their respective females in several other of the specimens studied (Fig. 6D, H, I). The only intact female primary appendage observed is shown in Fig. 6C attached laterally to the suprabasal cell of the receptacle of an immature individual from which the perithecial upgrowth is arising.

Discussion
When he described Rhizopodomyces, Thaxter (1931) did not have immature specimens of the type species, R . merragatae , available for study. The two mature females he observed lacked any vestige of an appendage and Thaxter believed that the perithecium of this species had a strictly terminal origin from the upper cell of a two-celled receptacle. The account presented here, which includes results of a study of juveniles of several species, shows that the perithecium of Rhizopodomyces spp. actually has an intercalary origin, arising laterally from the suprabasal cell of a two-celled receptacle that is derived from the lower cell of the bicellular spore. A simple, unicellular appendage, derived from the upper cell of the spore, is present during early stages of development of the female but it may more or less disintegrate or apparently be easily detached prior to or soon after maturation of the thallus. Enlargement of the suprabasal cell of the receptacle and the primary stalk cell of the perithecium coupled with their realignment into a more or less straight axis masks the original lateral origin of the perithecial primordium. Thaxter (1931) suggested that Rhizopodomyces resembles the dioecious Amorphomyces Thaxter (1893) most closely "in producing a strictly terminal perithecium and in the two-celled receptacle of the male individual.'' He also thought that Dioicomyces and three apparently allied dioecious genera, Dicrandromyces Thaxter (1931) , Triandromyces Thaxter (1931), andTetrandromyces Thaxter (1931) are related to Amorphomyces, although their females, unlike those of Amorphomyces bear a well-defined appendage. The females of Dicrandromyces, Triandromyces , and Tetrandromyces are nearly identical in basic design to those of Dioicomyces , and these genera were separated from one another by Thaxter on the basis of their males as will be discussed below.
Eleven species of Amorphomyces have been described , all from Staphylinidae (Coleoptera), and mostly on members of subfamily Aleocharinae. Twenty-six of the 32 described taxa of Dioicomyces occur on beetles of the family Anthicidae (Coleoptera), and the same group o{ insects provides the hosts for the known species of Dicrandromyces (1 species) and Triandromyces   on Colydiidae. Benjamin (1970) described D. mesoveliae, an anomalous species on a true bug of the family Mesoveliidae , and Balazuc (1971Balazuc ( , 1975 described D. bournieri on Phlaeothripidae (Thysanoptera), the only known laboulbenialean parasite ofthrips. The eight taxa of Rhizopodomyces discussed in this paper parasitize Hebridae (Hemiptera) , but it cannot be presumed that other species will not be discovered on members of other groups of insects that coinhabit the same environment as the hebrids. Although many genera of Laboulbeniales have apparently very limited host ranges, others have become adapted to a broad spectrum of insect groups and other arthropods (Benjamin 1967(Benjamin , 1971). Host range can be an uncertain guide in assessing relationships or in making taxonomic judgements in the Laboulbeniales.
Except for Dioicomyces mesoveliae in which the appendage of the female is two celled, the appendage of the female of all other species of Dioico-myces and of those of Dicrandromyces, Triandromyces, and Tetrandromyces is one celled as in Rhizopodomyces species. The appendage of all species of Dioicomyces and its presumed allies is persistent whereas in those of Rhizopodomyces it often deteriorates or apparently is readily broken off. Species of Amorphomyces differ from all others in the Laboulbeniales in that the mature thallus shows no evidence of having borne an appendage. The apparent lack of this structure in members of this genus was easily explained earlier (Thaxter 1931) because of a misconception regarding the nature of the spore. Thaxter always maintained that the spore of Amorphomyces spp. was unique in the Laboulbeniales in being nonseptate, and this idea was accepted without question until Tavares (1970) demonstrated that it is indeed two celled. She also described not only the lateral origin of the perithecium in this genus but also the development of a distinctive type of rudimentary two-celled appendage on the female individuals. This appendage can be easily detected only on very immature stained specimens. Its remnant is incorporated into the wall of the perithecium and cannot be readily observed on individuals in late stages of development and at maturity. The spore of Amorphomyces spp. differs from that of other genera in that its upper segment is very small compared to the lower segment. It might be mentioned in passing that Thaxter undoubtedly observed the spore septum in a species of Amorphomyces when he described A. obliqueseptatus in 1900. He later transferred the species to Dioicomyces as D. obliqueseptatus because of its septate spores. The species had been based on a few female specimens lacking intact receptacles. There was no evidence of an appendage on the females, and males were missing. Thaxter illustrated D. obliqueseptatus in the 1908 segment of his monograph (Pl. XLII,). His figure 17 shows a spore having an oblique septumexactly the kind of spore now known to occur in Amorphomyces spp.
The receptacle of the perithecium-bearing individual in many genera of Laboulbeniales, both monoecious and dioecious, may appear to be two celled, but a true two-celled receptacle is rare in the order and appears to be constant only in the females of Rhizopodomyces and Amorphomyces (Tavares 1970). Other genera with simple receptacles, like Dioicomyces, Dicrandromyces, Triandromyces, and Tetrandromyces, always have a third receptacular cell (Ill) lying more or less above cells I and II. In these genera cell III subtends the primary appendage, i.e., the modified upper spore segment, and all three cells of the receptacle, although often small, are well defined and clearly distinguished from the often elongate stalk cell of the perithecium which diverges from the suprabasal cell, i.e., cell II. In Amorphomyces the basal cell of the receptacle appears to comprise the whole of the receptacle proper, and it often is separated from the cells above by a slight constriction that may become more or less suffused. The small suprabasal cell of the receptacle, which subtends the rudimentary appendage, is intimately associated with the stalk cells and basal cells of the perithecium and is not at all conspicuous in mature individuals. The two-celled receptacle of the female of Rhizopodomyces is very different from that of Amorphomyces in being separated from the body of the perithecium and its basal cells by the elongate primary stalk cell. The suprabasal cell usually comprises the greater part of the receptacle of Rhizopodomyces spp. and bears a free, well-developed, unicellular appendage-at least during development of the thallus . Unlike Rhizopodomyces , the basal cell of the receptacle of species of Amorphomyces and Dioicomyces and its allies never develops rhizoidlike outgrowths and typically forms a blackened, indurated foot like that of most Laboulbeniales.
The males of the taxa here being discussed, although simple in structure and resembling one another superficially, display distinct differences from genus to genus. The lower cell of the male spore of species of Dioicomyces and Tetrandromyces remains undivided and forms a one-celled receptacle (Benjamin, unpublished). The upper spore segment typically becomes divided into three smallish cells by the formation of two cross walls. In Dioicomyces the terminal cell is converted directly into a single antheridium having either a terminal or subterminal discharge tube. A second antheridium forms only if the first aborts (Thaxter 1931). In Tetrandromyces the terminal cell of the four-celled thallus gives rise to two small cells disposed side by side each of which gives rise in turn to a pair of antheridia; the four antheridia thus formed are adnate at the base with only their discharge tubes free. The male of Tetrandromyces seems sufficiently distinct from that of Dioicomyces to warrant recognition of the genus.
In Triandromyces and Dicrandromyces the lower cell of the male spore divides and forms a two-celled receptacle (Thaxter 1931), the original spore septum being more or less oblique. The upper spore segment of Triandromyces typically divides and forms two cells; occasionally it remains undivided or sometimes may form three cells. The terminal cell of four-celled thalli usually gives rise to two free antheridia and the subterminal cell to one free antheridium; three-or five-celled thalli may give rise to only two or as many as five antheridia (Thaxter 1931). In Dicrandromyces the upper spore segment produces a pair of divergent cells from each of which arise two free antheridia. In their two-celled receptacles and the production of free antheridia, the males of Dicrandromyces and Triandromyces are distinctly different from those of Dioicomyces and Tetrandromyces. Because so few species of Dicrandromyces, Triandromyces, and Tetrandromyces are known it is perhaps premature, as suggested by Thaxter (1931), to stress the taxonomic significance of the male thallus at the generic level, but the question should be kept open .
The lower segment of the male spore of Amorphomyces spp. also becomes once septate and forms a simple two-celled receptacle while the initially small upper segment enlarges and is converted directly into a flaskshaped antheridium with an elongate, terminal or slightly eccentrically placed discharge tube (Tavares 1970;Thaxter 1931).
Like members of Dicrandromyces, Triandromyces, and Amorphomyces, the lower segment of the male spore of Rhizopodomyces spp. becomes once septate and forms a two-celled receptacle . The upper spore segment of some species, as in all species of Amorphomyces, forms a simple antheridium directly; however, in other species it becomes a simple, sterile appendage resembling that of the female individuals . In the latter species, the suprabasal cell of the receptacle divides and forms one or two antheridia directly and what amounts to a secondary suprabasal cell. Sterile appendages are unknown on the males of Amorphomyces, Dioicomyces , Dicrandromyces, Triandromyces , and Tetrandromyces, and in no instance has the male of any of these genera been found to form antheridia from a cell derived from its receptacle.
Perithecial ontogeny in Rhizopodomyces spp. is basically like that described by Thaxter in 1896 for Stigmatomyces baeri (Knoch) Peyritsch, Peyritschiella geminata Thaxt., and Laboulbenia flagella ta Peyritsch. Thaxter subsequently often referred to such perithecia as having "normal development," and the sequence of cell divisions accompanying formation of such perithecia is the one most commonly encountered in the Laboulbeniales (Benjamin 1971: 43). Briefly, in this pattern of development a branch arising from the receptacle becomes divided near the tip. The upper cell ultimately forms the female organ which gives rise to asci, and the lower cell gives rise to the enveloping perithecium consisting of two stalk cells, three basal cells, and four inner and outer vertical rows of wall cells. The number of cells in each row of outer wall cells may vary, being three to many, but most commonly it is four or five or a combination of these numbers. The number of cells may be constant in the species of some genera in which the number is relatively small according to Tavares and Majewski (1976). This has proved to be the case in the species of Rhizopodomyces observed thus far where there are four cells in three rows and five cells in the fourth row. Thaxter stated that there were five cells in each row of outer wall cells in Amorphomyces and Dioicomyces and implied that this was true also for Dicrandromyces, Triandromyces, and Tetrandromyces (Thaxter 1931), although he did admit to some doubt that this was true for all the rows . Tavares (1970) demonstrated the so-called "normal" pattern of early perithecial development in Amorphomyces falagriae Thaxt., the type species of the genus, but did not comment on the number of cells in each row of wall cells in the mature perithecium. I have examined immature and mature perithecia of A. falagriae and several species of Dioicomyces. In all instances where early development could be followed, I have found the " normal" sequence of cell divisions leading to wall cell formation in Dioicomyces. Outer and inner wall cell configurations in mature females of Amorphomyces and Dioicomyces are difficult to interpret, especially near the perithecial tip. Here, cellular boundaries often are obscured by mature spores and the persistent inner wall cells. The problem is compounded in Dioicomyces spp. where the spores usually begin to germinate prior to discharge as evidenced by differentiation and suffusion of the foot. In no instance, however, have I found five cells in all rows of outer wall cells in A. falagriae or in the species of Dioicomyces studied. In A. falagriae, where the rows of cells are somewhat spirally arranged, each row appears to consist of three elongate cells until maturation when the terminal cell of one row cuts off a fourth small cell. In all species of Dioicomyces examined, each row of outer wall cells consists of four cells until perithecia approach maturity. In D. mesoveliae there is no evidence of further division of any of the terminal cells, whereas in the other species the terminal cell of one row appears always to cut off a fifth cell. In a few instances I found what appears to be a fifth extremely small cell cut off terminally by the wall cell row opposite the five-celled row .
Structural changes in perithecia of Laboulbeniales by supplementary divisions of wall cells at or near the apices, formation of specialized outgrowths by wall cells, and alterations in perithecial shape undoubtedly are in some way correlated with ascospore discharge. Such modifications may affect opening and closing of the ostiole when perithecia contact a host. There possibly may be differences in the number of cells formed at the tip of the perithecium in different species of a given genus depending on adaptations relating to spore dispersal. Dioicomyces, a relatively large genus, exhibits considerable variation in perithecial structure, or at least conformation, from species to species. In some, the perithecium is little modified, being nearly straight or only slightly curved and with a relatively undistinguished tip (i.e., D. anthici, D. trinitatis Thaxt., D. Jormicellae Thaxt., D. mesoveliae, etc.). In a few species the perithecium itself may become greatly altered in shape as in D. umbonatus Thaxt. , D. falcatus Spegazzini, and especially D. malleolaris Thaxt. where the long axis of the ascus-containing part becomes oriented nearly at right angles to the primary stalk cell. In still other species, wall cells at or near the tip of the perithecium produce thickened projections which presumably function as trigger organs (i.e., D. spiniger Thaxt., D. glossophorus Speg., D. myrmecophilus, D. onchophorus Thaxt., D. prominens Thaxt. , D. rostellatus Speg.). Comparative studies of perithecial development in this genus are needed to determine what effect, if any, marked structural modifications may have on wall cell number, especially near the tip.
Only one species of Amorphomyces, A . biformis Thaxt. , forms any kind of specialized perithecial outgrowth. In this species Thaxter (193 I) described two types or forms (a and b) of females, both more or less crescent shaped; in form a the perithecium develops a conspicuous, elongate, median outgrowth projecting nearly at right angles to the long axis of the perithecium; form b develops no such outgrowth. Other species of Amorphomyces typically have perithecia that are nearly straight or only slightly curved.
Four species of Rhizopodomyces , R. merragatae , R. californicus, R. erect us, and R. basifurcatus, have perithecia in which the long axis is nearly in line with that of the primary stalk cell. It may be significant that in these species the terminal cell of the five-celled row of wall cells is more or less protrudent, which is not the case in the geniculate species, R. geniculatus, R. mexicanus , and R. polhemusii. One can speculate that in these two groups of species perithecial shape and the nature of the apical wall cells are correlated with spore discharge. It is unfortunate that mature perithecia are unknown in Rhizopodomyces sp. where extreme modification of one of the basal cells gives rise to an appendage of remarkable size but unknown function.
From the above considerations, it can be seen that species of Rhizopodomyces, Amorphomyces, Dioicomyces, Dicrandromyces , Triandromyces, and Tetrandromyces resemble one another in several ways: (1) All are dioecious and have simple males consisting of only a few cells. The male typically lacks an appendage and produces one or a small number of antheridia that are derived from the upper, smaller segment of the two-celled spore; the only exception is Rhizopodomyces where in a few species the upper spore segment forms a sterile appendage like that of the female and the antheridia arise from the suprabasal cell of the receptacle. (2) The receptacle of the male thallus consists of only one or two cells ; that of the female thallus consists of only two or three cells. (3) A simple primary appendage derived from the upper spore segment is borne by the female receptacle and consists of one , sometimes two , cells. (4) Except for Rhizopodomyces, in which there is some evidence in a few species that more than one perithecium may be produced , the female thallus typically gives rise to only one perithecium which always arises from the suprabasal cell of the receptacle. And except for Amorphomyces, the primary stalk cell of the perithecium usually is relatively large and frequently forms an elongate perithecial stipe. (5) The sequence of cellular divisions involved in perithecial formation is the same in all of the species and follows what is termed the " normal pattern of development" in the Laboulbeniales. The number of cells in each vertical row of wall cells is small, most commonly four or five, but as few as three in Amorphomyces .
In comparative morphological studies like the one presented here, one is tempted to speculate on possible interrelationships of the organisms examined. In the Laboulbeniales, as in most fungi where there is no fossil record, such speculations must be based on characteristics of extant species.
In the absence of physiological data, the primary characteristics available are similarities and differences in structure .
Laboulbeniales probably are monophyletic (Benjamin 1973;Denison and Carroll 1966). The group, however, must be of ancient origin, its evolution coinciding with that of its arthropod hosts, and the great diversity of generic types suggests that many divergent phylogenetic lines have developed at different times throughout its long evolutionary history , an evolution that undoubtedly is progressing rapidly in present time. It is to be expected that similarities in thallus structure reflect close relationship within some groups of genera as a result of parallel evolution from a common ancestor. An example of this would seem to be especially evident in a number of genera centered around Stigmatomyces Karsten (Benjamin 1970;Tavares 1973). However, similarities in structure also may result from convergent evolution and this phenomenon would lead to superficial resemblances of genera having distantly related ancestral types . I believe that Rhizopodomyces resembles Amorphomyces and Dioicomyces as a result of convergence and that the former genus is not closely related to the fatter genera.
Despite the fact that the perithecium-bearing individual in Amorphomyces and Rhizopodomyces has a two-celled receptacle-a characteristic known only in these two genera of Laboulbeniales-the genera are very distinct. The lack of cell III in the receptacle of Amorphomyces may be the result of reduction of a formerly three-celled receptacle, as is found in Dioicomyces , Dicrandromyces, Triandromyces , and Tetrandromyces, coupled with the incorporation of the suprabasal cell, appendage, and stalk cells along with the true basal cells into the functional base of the perithecium. Amorphomyces appears to be one of the most specialized genera of Laboulbeniales having simple receptacles . The two-celled receptacle bearing a well-defined , though somewhat ephemeral, simple appendage of Rhizopodomyces suggests that this genus may be relatively primitive and less specialized than Amorphomyces and also Dioicomyces , etc ., in which the appendage is persistent, often thick walled. I suspect that the receptacle of Rhizopodomyces is primitively two celled and that its precursor lacked a cell III. The foot of Rhizopodomyces spp. is less highly modified than in the latter genera and in the female part of the function of support if not attachment is assumed by lobate or rhizoidlike outgrowths of the basal cell.
Another possibly primitive characteristic of Rhizopodomyces is found in the males where a tendency for antheridia to be derived from the upper segment only of the spore has not become stabilized, as apparently it has in Amorphomyces and Dioicomyces and its allies . In some species of Rhizopodomyces , the suprabasal cell of the receptacle of the male , like that of the female , gives rise to the functional sexual organ , and the original terminal segment of the spore forms a sterile appendage. In this way the male of these species of Rhizopodomyces resembles the female, except for its sexual structures, as in Dimeromyces Thaxt. (1896) and its allies, Dimorphomyces Thaxt. (1893), Nycteromyces Thaxt. (1917), and Trenomyces Chatton & Picard (1908, 1909, and especially Laboulbenia formicarum Thaxt. (1902;Benjamin and Shanor 1950) which is perhaps one of the most primitively dioecious of all Laboulbeniales except possibly Herpomyces Thaxt. (1902). All species of Herpomyces, which occur on cockroaches, are dioecious, but sexual dimorphism may not be obligate in the genus. In her studies of H. paranensis Thaxt., Tavares (1965) confirmed an earlier observation of Thaxter (1908) that perithecial abortion on the female thallus of this species often leads to the formation of functional antheridia identical to those on the male thallus. Whether or not this is true of other species of Herpomyces still is unknown.
The male and female ascospores of some dioecious Laboulbeniales are conspicuously different in size, the female often much larger than the male, and spore dimorphism undoubtedly is a derived characteristic in the order. Ascospore dimorphism is much more pronounced in Amorphomyces and Dioicomyces than in Rhizopodomyces-where the male and female spores are nearly of equal size-and most dioecious Laboulbeniales with a few exceptions which will be discussed in a future paper.
In conclusion, I believe that Rhizopodomyces is not closely related to Amorphomyces or Dioicomyces and, like so many genera of Laboulbeniales, cannot be allied directly with any other of the known genera. Dioicomyces, Dicrandromyces, Triandromyces , and Tetrandromyces appear to be very closely related on the basis of the female thallus. If distinct differences in their male thalli prove to be stable as additional species, especially of the latter three genera, are discovered, then these genera should continue to be recognized as distinct. Amorphomyces appears to be related to Dioicomyces and its allies, but may represent the most highly specialized member of this complex of dioecious genera.