THE ANATOMICAL RECORD 248:176–188 (1997) Ultrastructure of the Parotid Salivary Glands in Seven Species of Fruit Bats in the Genus Artibeus BERNARD TANDLER,1* TOSHIKAZU NAGATO,2 AND CARLETON J. PHILLIPS3 1Department of Oral Anatomy II, Kyushu Dental College, Kitakyushu 803, Japan 22nd Department of Oral Anatomy, Fukuoka Dental College, Fukuoka 814, Japan 3Department of Biological Sciences, Illinois State University, Normal, Illinois ABSTRACT Background: In previous studies, we determined that the submandibular glands of five species of Neotropical fruit bats in the genus Artibeus had seromucous granules in their demilune cells with substructures that varied interspecifically in accordance with systematic relationships. Moreover, the striated ducts in these frugivores exhibited structural modifications that apparently are related to the consumption of a diet rich in potassium, but deficient in sodium. We now turn our attention to the parotid gland in a large number of species in this genus to determine if it follows the same structural pattern as does the submandibular gland. Methods: Members of seven different species of Artibeus were livetrapped in various Neotropical locations. The parotid glands were extirpated from euthanized bats, fixed in the field, and prepared for electron microscopic examination by conventional means. Results: The parotid glands in all seven species were virtually identical in morphology. The acinar cells (determined to be seromucous on the basis of ultrastructural criteria) contain large numbers of what appear to be vacuoles, but which are a type of secretory granule. These granules have an electronlucent matrix and may contain one or several circular membranous profiles arranged either concentrically or in a random array. These granules appear to form by progressive dilatation of the termini of Golgi saccules, with the nascent granules finally severing their connection with the Golgi apparatus. Many of the internal membranous profiles are formed simply by invaginations of the limiting membrane of the granule; others may result from indentation of the limiting membrane by protrusions from adjacent granules; the source of multiple internal membranes in certain granules is unclear. The exocytosis of these granules results in the acinar and intercalated duct lumina being filled with an abundance of membranous material. Such extruded membranes are present in some striated ducts, but not in others, suggesting that they are degraded during passage through the duct system. The striated ducts are of conventional appearance, lacking the frondose processes that are prominent in the submandibular glands of Artibeus. Conclusions: The parotid gland in Artibeus shows none of the interspecific ultrastructural variability that characterizes the submandibular gland in bats of this genus. The seromucous acinar cells secrete granules that release phospholipids as well as glycoconjugates into the saliva. Based on the lack of frondose processes with their sodium-transporting portasomes, the striated ducts of the parotid gland are less concerned with electrolyte homeostasis than are those in the submandibular gland. Anat. Rec. 248:176–188, 1997. r 1997 Wiley-Liss, Inc. Key words: parotid gland; salivary glands; seromucous cells; Golgi apparatus; secretion; striated ducts; chiropterans; fruit bats Salivary glands have had a major biological role in the evolution and adaptive radiation of mammals. In part, this is because these organs can produce products of immediate survival value, can be responsive to both r 1997 WILEY-LISS, INC. *Correspondence to: Dr. Bernard Tandler, Department of Oral Anatomy II, Kyushu Dental College, 2-6-1 Manazuru, Kokurakitaku, Kitakyushu 803, Japan. Received 4 December 1996; accepted 23 December 1996. 177 PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS external and internal environments, and can exhibit remarkable plasticity in terms of gene regulation and secretory processes (Phillips, 1996; Phillips and Tandler, 1996). In the present article, we describe the ultrastructure of the parotid salivary glands in seven species of Neotropical fruit bats in the genus Artibeus. In 1956, Wimsatt reported that the acinar cells in the Artibeus jamaicencis parotid gland are unusual at the light microscopic and histochemical levels (Wimsatt, 1956). It was later shown that the ultrastructure of these cells in A. phaeotis is also unusual in that the secretory granules look like vacuoles rather than typical acinar cell products (Phillips et al., 1977). Although the acinar secretory cells in the Artibeus parotid gland superficially resemble mucous cells, they actually are seromucous cells according to the criteria set forth by Young and van Lennep (1978) and by Tandler and Phillips (1993). Moreover, our previous comparative analyses of the bat family Phyllostomatidae have shown that, from an evolutionary perspective, the unusual acinar cells in Artibeus most likely evolved from a more typical parotid serous cell (Phillips et al., 1987, 1993). In the present comparative study of a number of Artibeus species, we sought to clarify the ultrastructure not only of the secretory cells, but of the remainder of the gland as well. Neotropical fruit bats of the genus Artibeus are geographically widespread, occurring from western Mexico through Central America and as far south as northernmost Argentina. They also have reached the Antilles and one species, A. jamaicencis, occurs on nearly every island (Phillips et al., 1989). For the present investigation, we selected seven species in this genus because we sought to determine whether or not there are interspecific differences in their parotid salivary glands, as there are in the submandibular glands (Tandler et al., 1986). The possibility of interspecific differences is important because the species selected represent two significantly different size categories (small, represented by A. cinereus, and large, represented by A. lituratus) with presumably different diets and different metabolic strategies. In addition, there has been some controversy about the selected species. Owen (1991) has argued that the species used in the present study represent three genera (Dermanura cinereus, Koopmania concolor, and Artibeus). However, recent molecular genetic analysis supports placement of all of the species in question in a common genus (Pumo et al., 1996). The case for inclusion of the species in a single genus would be strengthened if their parotid salivary glands were identical or similar. We previously have shown the systematic value of salivary gland ultrastructure (e.g., Tandler et al., 1986) and one would anticipate that A. cinereus and A. concolor would substantially differ from one another and from the other five species if they truly represent different genera. MATERIALS AND METHODS Parotid glands were obtained from seven species of fruit bats in the genus Artibeus. The bats were captured at night in mist nets as they were feeding or using fly-ways leading to and from fruit trees. Identification of some of the South American species in this genus can be difficult; in the present investigation, we used the criteria of Lim and Wilson (1993) for species identification. Voucher specimens used in this research were deposited in the research collection at the Carnegie Museum of Natural History, Pittsburgh. The species examined, sample sizes, geographic origins, and mode of initial fixation of parotid gland specimens are listed below: A. cinereus (n 5 7); Suriname; triple aldehyde-DMSO A. jamaicensis (n 5 4); Jamaica and Trinidad; Karnovsky-DMSO A. lituratus (n 5 3); Suriname; triple aldehyde-DMSO A. planirostris (n 5 3); French Guiana; KarnovskyDMSO A. concolor (n 5 1); Suriname; triple aldehyde-DMSO A. amplus (n 5 1); Suriname; triple aldehyde-DMSO A. obscurus (n 5 1); French Guiana; Karnovsky-DMSO After capture, bats were euthanized in the field with T-61 euthanasia solution and the parotid glands quickly removed and fixed by immersion in either triple aldehyde-DMSO (Phillips, 1985) or in half-strength Karnovsky’s (1965) fluid in 2.5% DMSO, a mixture that we formulated for field fixation of salivary glands. After transport to the laboratory, aldehyde-fixed tissues were thoroughly washed in buffered sucrose, then postfixed in phosphate-buffered 2% osmium tetroxide (Millonig, 1961a). After rinsing in distilled water, the specimens were soaked overnight in acidified 0.25% uranyl acetate (Tandler, 1990). Another rinse in distilled water was followed by dehydration in ascending concentrations of ethanol, passage through propylene oxide, and embedment in Epon:Maraglas (Tandler and Walter, 1977). Thin sections were stained with acidified methanolic uranyl acetate (Tandler, 1990), followed by staining with lead tartrate (Millonig, 1961b) or lead citrate (Venable and Coggeshall, 1965), and examined in a JEOL (Tokyo, Japan) 1200EX or Siemens (Knoxville, TN) Elmiskop 1A or 101 electron microscope. Semithin sections were stained with methylene blueazure II (Richardson et al., 1960) and examined in an Olympus Vanox (Tokyo, Japan). RESULTS Histologically, the parotid glands of Artibeus in all seven species match Wimsatt’s (1955, 1956) light microscopic descriptions based on paraffin sections; they consist of secretory endpieces and ducts in the usual arrangement (Tandler, 1993a), although the intercalated ducts are quite short. In semithin sections stained with methylene blue-azure II, the acinar cells appear to be filled with unstained vacuoles. The ducts are unremarkable. At the ultrastructural level, the parotid gland acini in all seven species are strikingly similar. At low magnification, the secretory cells [seromucoid according to histochemical tests (DiSanto 1960); seromucous according to ultrastructural criteria (Young and van Lennep, 1978; Tandler and Phillips, 1993)], echoing their light microscopical appearance, contain large numbers of putative vacuoles (Fig. 1). At higher magnification (Fig. 2), it becomes apparent that these structures are not vacuoles, but represent secretory granules of a peculiar type. In many granules, the matrix is totally ‘‘empty’’ 178 B. TANDLER ET AL. Fig. 1. Artibeus jamaicensis. Survey electron micrograph of a parotid gland. The acinar cells appear to be filled with vacuoles. The arrow indicates a distal intercalated duct that is shading into a striated duct. 31,500. (Fig. 3), but in some—not all—of the acinar cells in A. cinereus, the matrix contains a sparse, weblike array of faint threads, and an eccentrically placed, dense spherule may be present. An almost ubiquitous substructural feature of the lucent matrix in all seven species is membranes, most often of a circular configuration. Although these internal membranes are of the same thickness (,7 nm) as are the limiting membranes of the granules, the two membrane systems (limiting and internal) probably are of different composition, since the internal membranes occasionally become contaminated by particles during staining with lead tartrate, whereas the limiting membranes of the same granules never do (staining with lead citrate, however, frequently leads to deposition of particles on both types of membranes). Usually, the internal membranes are disposed as a single large circle, but may have several satellite circles, either inside or outside themselves. Some granules contain a number of more or less concentric circular membranes; this configuration is especially prominent in A. planirostris, where such a panoply may consist of well above a dozen (Fig. 4). Certain of the internal membranes may be the result of the invagination of a granule by a pseudopod-like extension of a neighbor (Figs. 5 and 6), but such a relationship should result in a pair of tightly apposed membranes (Fig. 7), a membranous profile rarely encountered in the Artibeus granules. At its simplest, the internal membranes appear to arise by invagination of the limiting membrane (Fig. 8), with the invagination eventually pinching off to form an independent circular membranous inclusion. Although in most of the species of Artibeus that we examined, the Golgi apparatus appears to consist of a number of separate dictyosomes, these seemingly independent structures in salivary gland secretory cells probably are all interconnected (Yamashina, 1995). In A. jamaicensis, the Golgi apparatuses appear to be in the form of small, independent dictyosomes. In fully charged seromucous cells in all seven species of Artibeus, the Golgi apparatuses are relatively inconspicuous. The Golgi saccules exhibit a degree of dilatation, with this expansion being most evident at the saccular terminations (Fig. 9). Although, in general, the seromucous cells have relatively few basal folds, a few acinar cells have an abundance of such folds and lateral folds as well (Fig. 10). Contiguous seromucous cells often form microvilluslined intercellular canaliculi (Fig. 11). That secretory granules engage in exocytosis into these canaliculi is PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS 179 Fig. 2. A. planirostris. Survey electron micrograph of a seromucous acinus. Echoing the previous figure, the cells appear to contain numerous vacuoles. 34,900. shown by the abundance of V-shaped invaginations along their limiting membranes (Fig. 12). Exocytotic granules occasionally are observed whose limiting membrane has fused with the plasma membrane and whose internal membranes are herniated into the acinar or intercellular canalicular lumen (Fig. 13). That the internal membranes are liberated during this process is shown by their abundance not only in the lumina of the acini (Fig. 14) and intercellular canaliculi, but in those of the ducts as well (see Fig. 16). The acinar cells exhibit all of the structural hallmarks of protein-secreting cells; as they become charged with secretory granules, the intervening cytoplasm is correspondingly reduced. In cells containing an abundance of seromucous granules, the mitochondria appear to have been displaced peripherally—a layer of these organelles appears to separate the surface folds from the rest of the cell. The intercalated ducts consist of simple cuboidal epithelium surrounding a lumen of relatively small caliber (Fig. 15) that often contains a plethora of circular membrane profiles. Each duct cell has a large, deeply notched nucleus and small aggregates of rough endoplasmic reticulum. The most noteworthy feature of these cells is the presence of a modicum of uniformly dense, serous-type secretory granules. These granules vary between 0.5–0.9 µm in diameter among the seven species of Artibeus. Lumina of proximal striated ducts contain circular membrane profiles (Fig. 16), whereas more distal ducts are devoid of such material (Fig. 17). Otherwise, the striated ducts are of typical appearance (Tandler, 1993c), with long, slender, basally situated mitochondria alternating with folded plasma membranes (Fig. 18). Mitochondria in the supranuclear cytoplasm tend to be ovate or elliptical. Short, flattened vesicles are fairly abundant in the region subjacent to the apical membrane, especially in A. lituratus, but the duct cells lack secretory granules altogether (Fig. 19). These cells bear variable numbers of short microvilli. Excretory ducts closely resemble striated ducts (Fig. 20), but are readily distinguishable from the latter by virtue of their mantle of fibrous connective tissue. The walls of these ducts harbor an occasional dark cell, 180 B. TANDLER ET AL. Fig. 3. A. planirostris. A seromucous granule that is totally devoid of substructure, appearing to be ‘‘empty.’’ 345,000. Fig. 4. A. planirostris. A seromucous granule that contains a multiplicity of more or less concentric membranes as well as some circular which is interlocked with the usual lighter cells (Fig. 20). It could not be determined whether these cells are fixation artefacts or whether they are a normal component of parotid excretory duct epithelium in a variety of Artibeus species. DISCUSSION The seromucous granules present in the parotid glands of the seven species of Artibeus that we examined are of a type rarely seen previously. Seromucous granules studied in other mammals tend to have a well-defined substructure, e.g., rat and mouse (Ciofi Luzzatto et al., 1968; Martinez-Hernandez et al., 1972), cat (König and Kühnel, 1986), mink (Tandler, 1991), ferret (Jacob and Poddar, 1987), and Siberian weasel (Matsunaga, 1992). In contrast, the matrix of the Artibeus granules is completely lucent. At first, the ultrastructural appearance of the parotid gland secretory granules of Artibeus seems to be a result of inadequate fixation. When osmium tetroxide was the mainstay of fixation for electron microscopy, secretory granules in a variety of salivary glands appeared extracted and empty (see, for example, Reinecke, 1967). With the advent of aldehyde fixatives, these same granules were seen to have a dense content (for comparison, see figure 40 in the compendium by Young and van Lennep, 1978). The obvious improvement in preservation of parotid secretory granule contents afforded by aldehyde fixation was documented by Robinovitch et al. (1966). The initial fixative that we used in our comparative ultrastructural studies, namely, triple aldehyde-DMSO (Kalt and Tandler, 1971, as modified by Phillips, 1985), seems to be a near universal fixative for salivary gland granules, preserving profiles inserted between the concentric membranes. Note that some of the internal membranes are contaminated by small, dense lead particles. 327,600. delicate details of granule substructure that permit differentiation of closely related species of bats (Tandler, 1993b; Tandler and Phillips, 1993). Moreover, in the case of the Artibeus glands, the same fixation protocol was applied simultaneously to the submandibular glands of the selfsame animals that furnished the parotid glands examined in the present study. The seromucous granules in the principal submandibular glands of four Artibeus species previously described (Phillips et al., 1977; Tandler et al., 1986) exhibited characteristic structural patterning for each individual species; the same holds true for the three additional species examined in the course of the present study (Tandler, unpublished observations). In other words, where a substructure existed in Artibeus submandibular gland seromucous granules, our fixation protocol revealed its presence. Furthermore, using the same protocol, we have not encountered granules of the Artibeus parotid gland type in the parotid glands of more than 230 other species of bats that we have fixed in the same manner. For these reasons, we believe that the ‘‘empty’’ appearance of the Artibeus parotid seromucous granules is representative of their true structure. The granules that most closely resemble those in the parotid glands of the fruit bats occur in standard pigs (Ferrandi, 1969; Boshell and Wilborn, 1978; Boshell, 1981). Boshell and Wilborn (1978) identify these granules as those of special serous cells, a cellular designation that we have argued elsewhere is based on inadequate fixation (Tandler and Phillips, 1987). Moreover, histochemical studies of the pig parotid gland have shown that the gland is seromucous in nature (Kamiya, 1977). Similar granules also are present in the parotid gland of Gottingen strain minipigs (Ginsbach and Küh- PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS Figs. 5–9. 181 182 B. TANDLER ET AL. nel, 1978; Lotz et al., 1990). Although the authors of the latter reports designate them as a mucous granules or as special serous granules, respectively, Ginsbach and Kühnel (1978) also mention that the minipig parotid gland secretes a ‘‘. . . watery, serous product . . .’’ leaving little doubt that the parotid secretory granules in this porcine variant also are seromucous in nature. The aforementioned studies on pig parotid glands were based on specimens initially fixed in glutaraldehyde, yet the granules maintained their lucency. Although it is true that certain secretory granules are better preserved by initial fixation in osmium tetroxide than in aldehyde mixtures (Tandler and MacCallum, 1974), this is not the case in the pig where, even after fixation solely in osmium, Yamada (1977) found the porcine parotid granules to have a lucent matrix. Finally, using morphological criteria (Young and van Lennep, 1978; Tandler and Phillips, 1993), we consider that their ultrastructural appearance indelibly marks pig parotid granules as seromucous. In addition to their lucency, another feature shared by seromucous granules in both Artibeus species and pigs is the presence of circular membranes in their matrices; these lack the tightly furled configuration of myelin figures (Revel et al., 1958). Although some of these membranes conceivably could result from the invagination of one granule by a process from another, some seromucous granules in the fruit bats have too many internal, circular membrane profiles (as exemplified by Fig. 4) to have originated in this fashion. Granules connected by a process usually have only a single membrane between the two moieties rather than two appressed membranes. Such a configuration is more in keeping with a fusion process (Tandler and Poulsen, 1976; Neutra and Schaeffer, 1977; Vidić, 1977; Tanaka, 1982; Shimono et al., 1984, Odajima and Nakane, 1984) than with internalization of limiting membranes of granules. We recently described mucous droplets in the accessory submandibular glands of long-winged bats (Miniopterus schreibersi and M. magnator) that were delimited by multiple membranes (Tandler et al., 1994); it was posited that these supernumerary membranes originated by the simple expedient Fig. 5.A. lituratus. These paired granules illustrate one possible way that some of the internal membranes originate. A protrusion from one granule penetrates a neighbor, invaginating its limiting membrane. A section passing through the plane indicated by the line would give rise to a granule with the configuration shown in the next micrograph. 340,000. Fig. 6.A. lituratus. A granule with two concentric internal membranes, conceivably resulting from a transversely-sectioned pseudopod within an invagination. 340,000. Fig. 7.A. lituratus. In these coupled granules, the surface protusion of the granule on the left has only a single membrane separating its contents from that of its consort. This configuration may be a prelude to granule fusion. 3121,000. Fig. 8.A. lituratus. This granule shows an early stage in the acquisition of internal membranes. The limiting membrane has formed a small invagination without the assistance of a neighboring granule. 321,000. Fig. 9.A. lituratus. A Golgi apparatus in an acinar cell fully charged with seromucous granules. The expanded portions of the saccules eventually may become separated from the Golgi membranes to form nascent secretory granules. 337,500. of Golgi saccules wrapping themselves around the nascent droplets. A similar mechanism cannot be operative in the Artibeus granules, because the membranes in question are suspended in their matrices, not applied to their surfaces. Figure 8 shows what probably is an early stage in the development of the internal membranes, with a surface membrane invaginating into the granule matrix to yield what in transverse section would appear to be a circular membrane profile. That granules conjoined by a pseudopod-like process are not simply produced by chemical fixation was shown by Dylewski and Keenan (1987), who used rapid freezing of rat mammary gland; ‘‘ball and socket’’ connections between granules were abundant. In order for the panoply of internal membranes that characterize some of the Artibeus parotid granules to have formed by this mechanism, the invagination process would have to be repeated many times. As the surface membranes become internal ones, they must undergo a physical or chemical change that leads to their becoming contaminated by lead stain when their precursors remain resistant to such an artifact. The origin of the lucent seromucous granules could not be unequivocally ascertained because we encountered virtually no secretory cells in any of our specimens that were at the beginning of their secretory cycle. It is in such cells that both the RER and Golgi apparatuses would be expected to be at their most prominent. The relatively small and inactive (on a morphological basis) Golgi apparatuses in cells filled with seromucous granules only hint at their role in producing these structures. At this point, the termini of the Golgi saccules are somewhat dilated and have a lucent content. Two possible Golgi-related pathways in the elaboration of the lucent granules present themselves. The simpler one is based on gradual enlargement of the termini of the Golgi saccules until the expansions attain the dimensions of mature granules, followed by scission of these amplified parts from the saccules to give direct rise to full-blown granules. A second possibility involves the release of relatively small granules from the termini of the Golgi saccules, and their progressive enlargement by accretion of progranules, a process that occurs in certain exocrine cells (Tooze et al., 1991; Lew et al., 1994). Regardless of which program the Golgi apparatus follows, the point at which the seromucous granules initially acquire their complement of internal membranes is undetermined. However, our observations suggest that this accession is a dynamic, ongoing process. The second of the two scenarios that we posited requires a massive reduction in surface area of the maturing granules (Lew et al., 1994); perhaps the internalization of membranes is an integral part of this process in seromucous granules of the Artibeus parotid type. The question arises as to what glycoproteins are present in the ‘‘empty’’-appearing granules of Artibeus. Of three different species of Artibeus tested by Junqueira et al. (1973) (A. concolor, A. jamaicensis, and A. lituratus), only A. jamaicensis had more than a trace of a-amylase, but the activity was trifling compared to that of bats such as the velvety free-tailed bat, Molossus ater, or the spear-nosed bat, Phyllostomus hastatus. A very low level of protease activity was evident in A. jamaicensis; the other two species had none. By histo- PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS Fig. 10. A. lituratus. The base of two seromucous acinar cells showing its array of basal folds. 317,600. Fig. 11. A. jamaicensis. A longitudinally sectioned intercellular canaliculus showing its close approach to the cell base. The canaliculus is lined with microvilli. 314,500. Fig. 12. A. lituratus. An intercellular canaliculus with connected membrane profiles (*) suggestive of a recent exocytotic event. 336,200. 183 Fig. 13. A. lituratus. A seromucous granule in the process of exocytosis, with an internal membrane herniating into an intercellular canaliculus. 356,200. Fig. 14. A. jamaicensis. An acinar lumen replete with numerous membranous vesicles and vacuoles representing exocytosed seromucous granule internal membranes. 311,100. 184 B. TANDLER ET AL. Fig. 15. A. jamaicensis. An intercalated duct. The cells contain a few secretory granules that have a maximum diameter of ,0.9 µm. 37,000. Fig. 16. A. lituratus. A longitudinally sectioned striated duct. Note the abundance of membranous detritus in the duct lumen. This material undoubtedly represents the internal membranes of the seromucous granules that were exocytosed into the acinar lumen and that so far have survived the exigencies posed by passage through the duct system. 33,000. Fig. 17. A. jamaicensis. A transversely sectioned striated duct whose lumen is devoid of membrane profiles. This sequence of membrane clearance from the lumina occurred in all seven species of Artibeus that were examined. 35,200. Fig. 18. A. jamaicensis. A portion of the striated duct wall showing the extremely long mitochondria alternating with folded plasma membranes. 38,400. PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS Fig. 19.A. lituratus. Numerous round and elongated vesicles are present in the supranuclear cytoplasm of a striated duct cell. 358,000. Fig. 20.A. jamaicensis. A portion of the wall of an excretory duct. The cells still retain a degree of basal striations. A dark cell is present at 185 the right in the ductular epithelium. It is not clear whether or not such cells are fixation artefacts, or represent a distinct population of duct cells. 39,500. 186 B. TANDLER ET AL. chemistry, the parotid acinar cells in A. jamaicensis were negative to antiserum against lysozyme (see figure 25 in Phillips et al., 1993). The expanding population of seromucous granules within acinar cells appears to ‘‘drive’’ the mitochondria toward the cell base where they form a layer that separates the basal folds from the rest of the cytoplasm. Unlike the situation in the mouse submandibular (Sharawy et al., 1978) and rat parotid (Sampson and Montalvo, 1983) glands, these organelles are not directly associated with the basal plasma membrane, and no energetic inferences should be drawn based on the propinquity of the two structures. The internal membranes of the mature seromucous granules are delivered into the acinar lumina during exocytosis of the seromucous granules, ultimately filling these passageways. Although pieces of plasma membrane are avulsed from salivary gland mucous cells during exocytosis (Tandler and Poulsen, 1976), such shed plasmalemmal fragments cannot by themselves account for the plethora of membranes present in the lumina of the fruit bat parotid glands. The circular membranes (which must be spherical in three dimensions) continue unaltered into the duct system. Whether some membranes survive transit through the main excretory duct to the mouth or while en route are degraded to their constituent molecules was not determined, although our observations suggest that breakdown of the membranes occurs in the striated duct. In either case, it is plain that a great deal of phospholipid is present in the final saliva entering the mouth. In the case of the two Miniopterus species mentioned earlier, we postulated that salivary lipids may be used to counter chemical defenses erected by insects that are their normal prey (Tandler et al., 1994). Phospholipidrich saliva might have a different role in species of Artibeus. In these frugivorous bats, the gastric glands have great numbers of acid-producing parietal cells and relatively few surface mucous cells or mucous neck cells (Phillips et al., 1984). The parietal cells in Artibeus [and in related phyllostomid fruit bats (Studholme et al., 1986)] ultrastructurally appear to be highly active and breakdown and passage of fruit is a remarkably rapid process. Studier et al. (1983) postulated that saliva may have a role in protection of the gastric mucosa, and Studholme et al. (1986) noted the possibility that surface-active phospholipids might be involved. Our documentation of salivary phospholipid in the form of membrane-like materials in the parotid initial saliva of Artibeus is consistent with hypotheses regarding mechanisms for protecting the stomach lining. The major difference between the striated ducts of the Artibeus parotid gland and those of the submandibular gland in the same animals is that the ducts in the former lack the array of leaflike microvilli (frondose processes) that characterize not only the latter, but striated ducts in the submandibular gland of most frugivorous bats (Tandler et al., 1989; Tandler, 1993c). Frondose processes have repeating units (portasomes) on the inner aspect of their limiting membranes that presumably function in sodium transport, i.e., Na1 conservation (Harvey et al., 1981), a necessity in fruit eaters, since their diet, although rich in potassium, is deficient in sodium. Whatever their ultimate function proves to be, the striated ducts in Artibeus parotid glands probably play a lesser role in electrolyte homeostasis than do those in their companion submandibular glands. Elsewhere, we have argued that salivary gland plasticity has been an extremely important factor in adaptive radiation in mammals (Phillips, 1996), and we also have shown that salivary glands in bats are greatly diversified, often at the species level, in obvious relationship to diet and other variables (Phillips and Tandler, 1987; Tandler et al., 1990; Phillips et al., 1993). By contrast, we have documented in the present comparative analysis an example in which seven different species of fruit bats of the genus Artibeus exhibit a ‘‘conserved’’ parotid gland, at least as far as secretory cell ultrastructure is concerned. An eighth species, A. phaeotis, can be added to the list on the basis of an earlier study (Phillips et al., 1987). This conservation is interesting because although in ultrastructural terms the parotid acinar cells are essentially the same in all eight studied species of Artibeus, this cell type differs from homologous cells in related genera of Neotropical fruit bats (Phillips et al., 1987; Tandler et al., 1986). Although there are some similarities among salivary glands in various genera of Neotropical fruit bats, the unusual secretory granule ultrastructure described in the present report thus is not a ‘‘signature’’ for frugivory in general. Instead, the granules described herein are a ‘‘generic’’ characteristic for just Artibeus and their conservation across the broad spectrum of the surveyed species of this genus implies that the parotid gland carries out a fundamentally important function in the lives of these related species. Salivary gland conservatism is not a general phenomenon in the genus, because the seromucous cells in the submandibular glands in Artibeus exhibit interspecific differences (Tandler et al., 1986). Generally speaking, the parotid gland and, more specifically, the parotid acinar cells appear to have two main roles in mammals. The obvious one is production and secretion of a variety of enzymes such as amylase and proteinases. However, as explained previously, this does not appear to be the case in Artibeus. At this point, it seems that phospholipids and some unknown but possibly nonenzymatic glycoprotein(s), perhaps prolinerich proteins, are the primary constituents of the watery parotid saliva. Another role of parotid acinar cells is production and release of HCO3, a function that has been extensively investigated in rats (e.g., Turner, 1993). Interestingly, the saliva of Artibeus is significantly more acidic and has a much lower buffering capacity than does the saliva of Carollia (a frugivore) or any of the tested insectivorous species (Dumont, 1996). The implication of these observations is that the parotid acinar cells in Artibeus do not buffer saliva to any meaningful extent and, to the contrary, might be different because they are exempt from this function. This would be consistent with the importance of high acid production by the gastric mucosa in Artibeus (Phillips et al., 1984). Patterns of diversity at the cellular level reflect evolutionary adaptation (Phillips, 1996). Patterns of conservation of sets of unique features are suggestive of ‘‘key innovations’’ (compare Nitecki, 1984) at the root or base of a group of species. In the case of the parotid gland in Artibeus, additional investigation is required in order to fully explain the biological significance of the PAROTID GLAND IN SEVEN SPECIES OF FRUIT BATS unusual parotid secretory granules in these particular bats. LITERATURE CITED Boshell, J.L. 1981 Effects of isoproterenol on the ultrastructure of pig parotid gland. Acta Anat., 109:270–274. Boshell, J.L., and W.H. Wilborn 1978 Histology and ultrastructure of the pig parotid gland. Am. J. Anat., 152:447–466. Ciofi Luzzatto, A., G. Procicchiani, and G. Rosati 1968 Rat submaxillary gland: An electron microscope study of the secretory granules of the acinus. J. Ultrastruct. Res., 22:185–194. DiSanto, P.E. 1960 Anatomy and histochemistry of the salivary glands of the vampire bat, Desmodus rotundus murinus. J. Morphol., 106:301–336. Dumont, E.R. 1996 Salivary pH and buffering capacity in fruit and insect-feeding bats. J. Mammal. (In press). Dylewski, D.P., and T.W. Keenan 1987 Ball and socket figures between secretory vesicles in mammary epithelial cells. Protoplasma, 138:65–68. Ferrandi, B. 1969 Prime osservazioni sulla fine struttura della parotide di maiale. Clin. Veterin., 92:407–417. Ginsbach, G., and W. Kühnel 1978 On the ultrastructure of the parotid gland in miniature pigs. Anat. Embryol., 155:23–36. Harvey, W.R., M. Cioffi, and M.G. Wolfersberger 1981 Portasomes as coupling factors in active ion transport and oxidative phosphorylation. Am. Zool., 21:775–791. Jacob, S., and S. Poddar 1987 Ultrastructure of the ferret parotid gland. J. Anat., 152:37–45. Junqueira, L.C.U., A.M.S. Toledo, and A.I. Doine 1973 Digestive enzymes in the parotid and submandibular glands of mammals. An. Acad. Brasil. Ciênc., 45:629–643. Kalt, M.R., and B. Tandler 1971 A study of fixation of early amphibian embryos for electron microscopy. J. Ultrastruct. Res., 36:633–645. Kamiya, N. 1977 Structure and carbohydrate histochemistry on the major salivary glands of the pig. Aichi-Gakuin J. Dent. Sci., 15:107–118 (in Japanese with English summary). Karnovsky, M.J. 1965 A formaldehyde-glutaraldehyde fixative of high osmolality for use in electron microscopy. J. Cell Biol., 27:137a– 138a (abstr). König, Jr., B., and W. Kühnel 1986 Licht- und elektronenmikroskopische Untersuchungen an der Glandula parotis und der Glandula submandibularis der Hauskatze. Z. Mikrosk. Anat. Forsch., 100:469–483. Lew, S., I. Hammel, and S.J. Galli 1994 Cytoplasmic granule formation in mouse pancreatic acinar cells. Evidence for formation of immature granules (condensing vacuoles) by aggregation and fusion of progranules of unit size, and for reductions in membrane surface area and immature granule volume during granule maturation. Cell Tissue Res., 278:327–336. Lim, B.K., and D.E. Wilson 1993 Taxonomic status of Artibeus amplus (Chiroptera: Phyllostomatidae) in northern South America. J. Mammal., 74:763–768. Lotz, S., J. Caselitz, H. Tschakert, W. Rehpenning, and G. Seifert 1990 Radioprotection of minipig salivary glands by orciprenalinecarbachol. Virchows Arch. Pathol. Anat., 417:119–128. Martinez-Hernandez, A., P.K. Nakane, and G.B. Pierce 1972 The secretory granules of the acinar cells of the mouse submaxillary gland. Am. J. Anat., 133:259–267. Matsunaga, S. 1992 Light and electron microscopic and histochemical studies on the major salivary glands of the Siberian weasel (Carnivora). Aichi-Gakuin J. Dent. Sci., 30:241–272 (in Japanese with English summary). Millonig, G. 1961a Advantages of a phosphate buffer for OsO4 solutions in fixation. J. Appl. Phys., 32:1636 (abstr). Millonig, G. 1961b A modified procedure for lead staining of thin sections. J. Biophys. Biochem. Cytol., 11:736–739. Neutra, M.R., and S.F. Schaeffer 1977 Membrane interactions between adjacent mucous secretion granules. J. Cell Biol., 74:983– 991. Nitecki, M.H. 1984 Evolutionary Innovations. University of Chicago Press, Chicago. Odajima, M., and F. Nakane 1984 Fusion and discharge of rat submandibular gland acinar secretory granules captured by quick freezing. Jpn. J. Oral Biol., 26:249–258. Owen, R.D. 1991 The systematic status of Dermanura concolor (Peters, 1865) (Chiroptera: Phyllostomatidae), with description of a new genus. Bull. Am. Mus. Nat. Hist., 206:18–25. Phillips, C.J. 1985 Field fixation and storage of museum tissue collections suitable for electron microscopy. Acta Zool. Fennica, 170:87–90. 187 Phillips, C.J. 1996 Cells, molecules, and adaptive radiation in mammals. In: Contributions in Mammalogy: A Memorial Volume Honoring Dr. J. Knox Jones, Jr. R.J. Baker and H.H. Genoways, eds. Mus. Texas Tech. Univ., Lubbock, pp. 1–24. Phillips, C.J., and B. Tandler 1987 Mammalian evolution at the cellular level. Curr. Mammal. 1:1–66. Phillips, C.J., and B. Tandler 1996 Salivary glands, cellular evolution, and adaptive radiation in mammals. Eur. J. Morphol., 34:155– 162. Phillips, C.J., G.W. Grimes, and G.L. Forman 1977 Oral biology. In: Biology of Bats of the New World Family Phyllostomatidae, Part II. R.J. Baker, J.K. Jones, Jr., and D.C. Carter, eds. Texas Tech University Press, Lubbock, pp. 121–246. Phillips, C.J., K.M. Studholme, and G.L. Forman 1984 Results of the Alcoa Foundation Suriname Expeditions. VIII. Comparative ultrastructure of the gastric mucosae in four genera of bats (Mammalia: Chiroptera), with comments on gastric evolution. Ann. Carnegie Mus. Nat. Hist., 53:71–117. Phillips, C.J., T. Nagato, and B. Tandler 1987 Comparative ultrastructure and evolutionary patterns of acinar secretory product of parotid salivary glands in Neotropical bats. In: Studies in Neotropical Mammalogy: Essays in Honor of Philip Hershkovitz. B.D. Patterson and R.M. Timm, eds. Field Mus. Nat. Hist., Chicago, pp. 213–229. Phillips, C.J., D.E. Pumo, H.H. Genoways, and P.E. Ray 1989 Caribbean island zoogeography: A new approach using mitochondrial DNA to study Neotropical bats. In: Biogeography of the West Indies: Past, Present, and Future. C.A. Woods, ed. Sandhill Crane Press, Gainesville, FL, pp. 661–684. Phillips, C.J., B. Tandler, and T. Nagato 1993 Evolutionary divergence of salivary gland acinar cells: A format for understanding molecular evolution. In: Biology of the Salivary Glands. K. DobrosielskiVergona, ed. CRC Press, Boca Raton, pp. 39–80. Pumo, D.E., I. Kim, C.J. Phillips, and H.H. Genoways 1996 Molecular systematics of the fruit bat, Artibeus jamaicensis: Origin of an unusual island population. J. Mammal., 77:491–503. Reinecke, M. 1967 Sekretionsstudien an der Acinuszellen der Parotis des Meerschweinchens im normalen Zustand, im Hunger und nach Pilocarpingabe. Arch. Klin. Exp. Ohr. Nas. Kehlkopfheilk., 189:168–179. Revel, J.-P., S. Ito, and D.W. Fawcett 1958 Electron micrographs of myelin figures of phospholipid simulating intracellular membranes. J. Biophys. Biochem. Cytol., 4:495–498. Richardson, K.C., L. Jarret, and E.M. Fink 1960 Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol., 35:313–323. Robinovitch, M.R., L.M. Sreebny, and E.A. Smuckler 1966 Preservation of the secretory granules of rat parotid gland for electron microscopy. Exp. Cell Res., 42:634–639. Sampson, H.W., and A.M. Montalvo 1983 Association of mitochondria with the plasma membrane of the acinar cells of actively secreting and nonsecreting rat lacrimal and parotid glands. Acta Anat., 116:353–357. Sharawy, M., G. Schuster, D. Pashley, and P. Palmer 1978 Association of mitochondria with the plasma membrane of the acinar cells of the mouse submaxillary glands. Acta Anat., 102:440–446. Shimono, M., K. Ashida, H. Yamane, Y. Iguchi, T. Horiuchi, S. Hashimoto, Y. Tanaka, and T. Yamamura 1984 Secretion and membrane fusion in the salivary gland. A study with freezefracturing and thin sectioning. Bull. Tokyo Dent. Coll., 25:177– 196. Studholme, K.M., C.J. Phillips, and G.L. Forman 1986 Results of the Alcoa Foundation Suriname Expeditions. X. Patterns of cellular divergence and evolution in the gastric mucosa of two genera of phyllostomid bats, Trachops and Chiroderma. Ann. Carnegie Mus. Nat. Hist., 55:207–235. Studier, E.H., B.C. Boyd, A.T. Feldman, R.W. Dapson, and D.E. Wilson 1983 Renal function in the Neotropical bat, Artibeus jamaicensis. Comp. Biochem. Physiol., 74A:199–209. Tanaka, Y. 1982 Membrane fusion in exocytosis of the rat parotid gland. A study with freeze-fracture. Shikwa Gakuho, 82:1083– 1108 (in Japanese with English summary). Tandler, B. 1990 Improved uranyl acetate staining for electron microscopy. J. Electron Microsc. Tech., 16:81–82. Tandler, B. 1991 Ultrastructure of the mink parotid gland. J. Anat., 177:185–194. Tandler, B. 1993a Introduction to mammalian salivary glands. Microsc. Res. Tech., 26:1–4. 188 B. TANDLER ET AL. Tandler, B. 1993b Structure of mucous cells in salivary glands. Microsc. Res. Tech., 26:49–56. Tandler, B. 1993c Structure of the duct system in mammalian major salivary glands. Microsc. Res. Tech., 26:57–74. Tandler, B., and D.K. MacCallum 1974 Ultrastructure and histochemistry of the submandibular gland of the European hedgehog, Erinaceous europaeus L. II. Intercalated ducts and granular striated ducts. J. Anat., 117:117–131. Tandler, B., and C.J. Phillips 1987 Ultrastructure of serous granules in the submandibular and parotid glands of the mouse opossum. Anat. Rec. 217:135A (abstr). Tandler, B., and C.J. Phillips 1993 Structure of serous cells in salivary glands. Microsc. Res. Tech., 26:32–48. Tandler, B., and J.H. Poulsen 1976 Fusion of the envelope of mucous droplets with the luminal plasma membrane in acinar cells of the cat submandibular gland. J. Cell Biol., 68:775–881. Tandler, B., and R.J. Walter 1977 Epon-Maraglas embedment for electron microscopy. Stain Technol., 52:238–239. Tandler, B., T. Nagato, and C.J. Phillips 1986 Systematic implications of comparative ultrastructure of secretory acini in the submandibular salivary gland of Artibeus (Chiroptera: Phyllostomidae). J. Mammal., 67:81–90. Tandler, B., C.J. Phillips, K. Toyoshima, and T. Nagato 1989 Comparative studies of the striated ducts of mammalian salivary glands. In: Cells and Tissues: A Three Dimensional Approach by Modern Techniques in Microscopy. P.M. Motta, ed. Alan R. Liss, Inc., New York, pp. 24–248. Tandler, B., C.J. Phillips, T. Nagato, and K. Toyoshima 1990 Ultrastructural diversity in chiropteran salivary glands. In: Ultrastructure of the Extraparietal Glands of the Alimentary Tract. A. Riva and P. Motta, eds. Kluwer Academic Publishers, Boston, pp. 31–52. Tandler, B., C.J. Phillips, and C.A. Pinkstaff 1994 Mucous droplets with multiple membranes in the accessory submandibular gland of long-winged bats. Anat. Rec., 240:178–188. Tooze, S.A., T. Flatmark, J. Tooze, and W.B. Huttner 1991 Characterization of the immature secretory granule, an intermediate in granule biogenesis. J. Cell Biol., 115:1491–1503. Turner, R.J. 1993 Ion transport related to fluid secretion in salivary glands. In: Biology of the Salivary Glands. K. DobrosielskiVergona, ed. CRC Press, Boca Raton, pp. 105–127. Venable, J.H., and R. Coggeshall 1965 A simplified lead citrate stain for use in electron microscopy. J. Cell Biol., 25:407–408. Vidić, B. 1977 Ultrastructural observations on the interaction of phospholipid bilayers during coalescence of secretory granules. Cytobiologie, 15:116–125. Wimsatt, W.A. 1955 On the nature of the interscapular glands of the tropical American fruit bat Artibeus jamaicensis Leach. Anat. Rec., 121:549–564. Wimsatt, W.A. 1956 Histological and histochemical observations on the parotid, submaxillary and sublingual glands of the tropical American fruit bat Artibeus jamaicensis Leach. J. Morphol., 99:169–210. Yamada, K. 1977 Ultrastructural diversity in mammalian parotid glands. Igaku Kenkyu, 47:93–114 (in Japanese with English summary). Yamashina, S. 1995 Dynamic structure and function of Golgi apparatus in the salivary acinar cells. J. Electron Microsc., 44:124–134. Young, J.A., and E.W. van Lennep 1978 The Morphology of Salivary Glands. Academic Press, London, pp. 26–28.