Microsmatic primates: Reconsidering how and when size matters

Abstract

The terms “microsmatic” and “macrosmatic” refer to species with lesser or greater levels, respectively, of olfactory function. Historically, primates are considered microsmats (olfactory sense reduced) with a concomitant increased emphasis on vision. The olfactory bulbs (forebrain centers that receive peripheral olfactory input) are proportionately smaller in primates compared to most other mammals. Similarly, the regions of the nasal cavity that are covered with olfactory epithelium (containing receptor cells) have proportionately less surface area in primates than other mammals. Thus, the generalization that primates are microsmatic is most frequently stated in terms of the proportional rather than absolute size of olfactory structures. Yet the importance of scaling to body size is unclear in regard to the chemical senses such as the olfactory or vomeronasal systems—do chemosensory structures such as olfactory bulbs and olfactory epithelium exhibit the same neural relationship to body mass that is seen for neural tissues that supply innervation to musculature or the skin? Previous studies examining neuronal density, volume, and/or surface area of the olfactory epithelium illustrate that different conclusions may be supported based on the parameter used. Plots of olfactory bulb volume versus body mass that generated for large-scale taxonomic studies or growth studies benefit from body mass (or total brain volume) with a comparative perspective. However, our examination of proportional versus absolute measurements implies that in comparisons within taxa, body size adjustments needlessly distort the data. As a final consideration, another embryonic derivative of the nasal placode, the vomeronasal organ, may warrant consideration regarding a definition of microsomia versus macrosomia. Anat Rec (Part B: New Anat) 279B:24–31, 2004. © 2004 Wiley-Liss, Inc.

INTRODUCTION

The primitive mammalian nasal chamber possesses two chemosensory neuroepithelial types that function as receptor organs for the main olfactory and accessory olfactory (vomeronasal) systems (Evans, 2003). These neuroepithelia are distributed across the nasal septum and turbinates and throughout the vomeronasal organ (VNO), respectively (Negus, 1958). Each derives from the same embryonic tissue found in the nasal placode and projects axons toward discrete portions of the forebrain (olfactory bulb and the more caudal accessory olfactory bulb). In monotremes and marsupials, both systems are highly developed (Wysocki, 1979). In some placental mammals, one or both of these systems appear to be in varied states of reduction (Bhatnagar and Meisami, 1998; Smith et al., 2001a). Olfactory retrogression may be clear in anosmatic mammals (e.g., toothed whales), where the olfactory system is completely lost. In other mammals, the degree of reduction may seem exaggerated by the elaboration of other portions of the brain (Martin, 1990).

The terms “microsmatic” and “macrosmatic” refer to differing levels of olfactory function among animals and were contrasted first to anosmatic groups by Turner (1891). Negus (1958) defined macrosmatic as “keen-scented” when referring to carnivores and grouped a range of mammals under the descriptor “microsmatic” (primates, seals, mysticete whales). Historically and to the present date, primates have been considered microsmats with a concomitant increased emphasis on visual sense (Elliot Smith, 1927; Zhang and Webb, 2003). The olfactory bulbs (regions of the brain that first receive peripheral olfactory impulses) are proportionately smaller in primates compared to most other mammals (Baron et al., 1983). Similarly, the regions of the nasal cavity that are covered with olfactory neuroepithelium have proportionately less surface area in primates than other mammals (Le Gros Clark, 1959). It is posited that the olfactory sense is further reduced in anthropoid primates than strepsirrhines (Box 1) and in diurnal primates compared to nocturnal primates (Martin, 1990). [Recently, much evidence has supported the division of primates into the suborders Haplorhini (monkeys, apes, humans, and tarsiers) and strepsirrhines (lemurs and lorises). Due to many unique features of tarsiers, we are excluding them from other haplorhines by using the term “anthropoid,” which is still considered a valid cladistic grouping to denote the close phyletic relationships of monkeys, apes, and humans (Fleagle, 1999)].

These generalizations are most strongly supported by the relationship between olfactory bulb volume and body mass (Stephan et al., 1981). A common practice in studies of broad ranges of taxa is to plot olfactory bulb volume against body mass (Box 1) (Stephan et al., 1981; Baron et al., 1983; Frahm, 1985). However, smaller-scale comparisons (interspecific or intersex) have been inconsistent, with some authors scaling chemosensory structures to body mass or total brain volume (Radinsky, 1979; Weiler et al., 1999; Mirich et al., 2002) and others comparing absolute measurements (Bhatnagar, 1975; Segovia and Guillamón, 1982) or both (Bhatnagar and Kallen, 1974, 1975; Smith et al., 2001b). This inconsistency in technique reflects different opinions on the best comparative approach. Several authors debated the need for scaling peripheral elements of the main olfactory and accessory olfactory systems when making comparisons between species or sexes (Dawley, 1998; Dawley et al., 1999; Farbman, 2000; Maico et al., 2003).

Box 1: Diversity in the Size of the Olfactory Bulb in Primates

Relative to the rest of the brain or body size, olfactory bulbs are proportionately smaller in strepsirrhines (Image 1A) compared to monkeys or apes (Image 1B). In absolute size, they may well measure the same volume. How is the difference in function best quantified? The interpretation is quite complex if olfactory sensitivity relates strongly to the number of receptor cells. First, the primary receptor cells are found in the nasal cavity (not in the olfactory bulb). Second, the olfactory bulb itself is a complex structure that initiates the process of odorant recognition and relays the neural impulses to other parts of the brain (Image 2). In a comparison of regression plots for all primates [data from Baron et al. (1983) and Stephan et al. (1981)], the relationship of olfactory bulb volume to total brain volume (2A) is more positive (i.e., the linear regression line slope is greater) than the relationship between olfactory bulb volume and body mass (Image 2B).

1. Image 1: Illustrations of a strepsirrhine (A) versus an ape (B) emphasizing that the olfactory bulbs (part of brain shown in solid black) may be of similar absolute size, but are proportionately larger compared to body size or compared to brain size (stippled) in strepsirrhines.

2A. Image 2A: log olfactory bulb volume versus log body mass for all representatives of major primate taxa. Image 2B: log olfactory bulb volume versus log total brain volume for all representatives of major primate taxa [data from Baron et al. (1983) and Stephan et al. (1981)].

Somatic afferent or efferent brain centers exhibit a relatively more clear relationship to body mass that relates to sensory or motor functions (Gahr, 1994). As body mass increases, greater muscle mass or surface area of skin requires a greater corresponding mass of neural input. Thus, a strongly positive relationship between body mass and neural mass necessarily implies that animals of disparate sizes will have regions of the brain that vary in tandem with body size, and functional questions are then framed in regard to proportional differences. Does this same caveat apply when examining peripheral and central elements of the chemical senses, such as the olfactory or accessory olfactory systems?

Chemosensory function may be measured physiologically in regard to discrimination and sensitivity. Sensitivity (the ability of concentration on odorant molecules) bears a strong relationship to receptor neuron population size (Negus, 1958; Meisami, 1989), whereas discrimination (ability to tell different odors apart) may relate more closely to the variety of receptor types or genes involved (McClintock, 2000). The inability to address discrimination and the finding that there are variable accumulations of nonfunctional olfactory genes (Rouquier et al., 2000) create an inherent weakness in quantitative studies of the primate olfactory bulb or olfactory epithelium. Quantitative studies can address sensitivity to the extent that it relates in large part to the overall number of receptors (Meisami, 1989; Meisami et al., 1990). Volume has been used in comparisons as a substitute (estimate) of receptor populations in chemosensory epithelia (Dawley, 1998). Thus, olfactory bulb or olfactory epithelial volume has been used to interpret differences in olfactory sensitivity between species. But should data be scaled to body size for such comparisons?

Olfactory bulb or olfactory epithelial volume has been used to interpret differences in olfactory sensitivity between species. But should data be scaled to body size for such comparisons?

If a particular strepsirrhine and a monkey or ape have olfactory bulbs of similar absolute sizes [true in several comparisons from Stephan et al. (1981)], does it make a functional difference that they are enormously different in body mass? Depending on the answer, our morphological approach to interpret that primates are microsmatic may have a weak foundation. Further scrutiny indeed may affirm that primates have relatively reduced olfactory sensitivity among mammals in general, and indeed there are other lines of evidence supporting this supposition (Rouquier et al., 2000). However, specific generalizations, for example, that nocturnal primates have a greater emphasis on olfaction compared to diurnal species (Martin, 1990), will remain far less certain unless based on firm morphologic criteria.

OLFACTORY EPITHELIUM

The olfactory epithelium is the site where the peripheral receptor cells (neurons) are located. These neurons possess a family of about 1,000 genes that probably encode for detection of various odors (Mombaerts, 1999). Neural impulses from the olfactory epithelium are transmitted to the olfactory bulb, a complex structure best examined cross-sectionally (Shepherd, 1972). Interpreting the relevance of body size to comparative studies is most intuitive when examining this peripheral receptor neuron population, in that its organization is simpler and thus more easily subjected to measurement via neuronal counts (Fig. 1). Yet there is little agreement or consistency in terms of approach to comparing olfactory epithelia between species. Quantitative studies of olfactory epithelia (Bhatnagar and Kallen, 1975) have measured volume, surface area, cross-sectional area, and receptor neuron measurements (density, nuclear diameter). Negus (1958) debated whether to place greater significance on absolute or proportional values when comparing olfactory epithelium area between humans and cats. Moulton and Beidler (1967) reviewed data from multiple sources and noted that, paradoxically, humans (microsmats) have similar olfactory epithelial surface areas as rabbits (macrosmats). In the same article, the authors later offered an apparent contradiction when calculating a great disparity in receptor neuron numbers between humans and rabbits. Moulton and Beidler (1967: p. 3) were aware that “no simple correlation exists between area of the olfactory epithelium” and olfactory function. In other words, area measurements do not necessarily covary with olfactory neuron numbers.

Figure 1. Micrographs of olfactory epithelium covering a turbinate in neonates of the mouse lemur (Microcebus murinus; A) and the mongoose lemur (Eulemur mongoz; B), and adults of the common marmoset (Callithrix jacchus; C) and greater bushbaby (Otolemur garnetti; D). Each image includes a 25 × 50 μm open rectangle for qualitative comparison of the density of nuclei.

Published data on insectivores revealed that species of Sorex (shrews) that vary by nearly 100% in body size can nonetheless have similar olfactory surface areas (LaRochelle and Baron, 1989). However, the olfactory receptor cell number estimates are much more similar between species due to differing cell densities (Table 1). This suggests that when inferring functional differences in the olfactory epithelium based on cross-sectional area or volume, differences in olfactory receptor cell size may be a critical variable. Body size in this example appears to be far less meaningful.

Table 1. Quantifiers of the body size and olfactory structures in insectivores*

	Body Mass (g)	Olfactory Epithelial Area (mm^2)	Olfactory Receptor Neuron Density (per mm^2)	No. of Olfactory Receptor Neuronsa	Main Olfactory Bulb Volume (mm^3)
* Compiled from Baron et al. (1983) and LaRochelle and Baron (1989). a Calculation of total olfactory receptor neuron population per nasal fossae (after adjustment for split error).
Blarina brevicauda	18.01	172.6	29,000	13,500,000	29.4
Sorex cinereus	5.01	74.5	39,000	6,500,000	11.6
S. fumeus	9.15	66.7	39,000	7,300,000	15.1
S. palustris	9.88	76.9	33,000	5,900,000	—

To elucidate further, Figure 1 shows olfactory epithelia from a diverse group of primates in terms of age (A and B, neonates; C and D, adults), activity patterns (nocturnal, diurnal, or active various times of day and night), and adult body size (6.5 g to 1.3 kg). The images in Figure 1 are arranged according to increasing body size, and each has a superimposed box of identical size. The receptor neuron bodies appear to be most densely packed, and nuclear diameter is smallest in the neonatal mouse lemur (Fig. 1A). Therefore, the same unit area of each tissue probably does not have a similar number of receptor neurons as in the other species. In a neonatal primate of larger body size (mongoose lemur; Fig. 1B), nuclear diameter appears to be larger, and correspondingly there are fewer neurons per unit area. However, there is little apparent difference in nuclear diameter between the neonatal mongoose lemur and an adult bushbaby (Fig. 1D); nuclear diameter appears to be largest in the smaller of the two adult primates shown (common marmoset; Fig. 1C). Density of receptor neurons appears to differ little between the two adult primates that differ in body size by as much as 400% (Fig. 1C and D). Of course, this example does not account for other variables that relate to total receptor neuron numbers. Yet, for purposes of discussion, it is clear that if the four species have an identical total olfactory epithelial area, the mouse lemur would have more receptors overall than the other neonate; the two adult primates would have similar receptor population numbers despite differing greatly in body size. This illustration might resolve the apparent contradiction that the microsmatic human has similar olfactory epithelial areas as compared to some macrosmats. Surface area does not account for differences in epithelial depth or volume or cell size and therefore cannot serve as a proxy for receptor numbers.

Anthropologists have not ignored this dilemma. Ankel-Simons (2000: p. 146) wrote, “It is not the size of the nasal membrane surface that determines acuity of the sense of smell, but the number and capacity of neural olfactory receptors per unit area.” If sensitivity correlates with receptor neuron numbers, then the above examples suggest that great care must be taken to control for similar cell densities. Comparing olfactory epithelium as percentages of total mucosal surface area continues to the present date (Ding and Dahl, 2003). Yet it is not clear how each type of nasal epithelial tissue scales to body size. Furthermore, scaling to body size or even nasal measurements, a practice sometimes used on the vomeronasal organ, (Weiler et al., 1999; Smith et al., 2001b), may severely distort volume or area data.

OLFACTORY BULB

Regarding the olfactory bulb, Martin (1990: p. 397) suggested “it is possible that development of other areas of the brain may simply overshadow the olfactory system in primate evolution.” Nonetheless, differences in olfactory bulb size among primates (e.g., strepsirrhines versus anthropoids) continue to be stated in proportional terms. Unfortunately, it is unclear whether volumetric differences in olfactory bulbs reflect differing numbers of neurons, supporting cells, or axons converging on elements of secondary neural relays. Six strata (Fig. 2) comprise the bulb and these may surround a central fluid-filled cavity (the olfactory ventricle); this morphology is analogous to the layers found in the cerebral cortex. From outside to center, the layers include primary olfactory axons (projections from the olfactory epithelium) and five additional internal layers that are involved in projections to and from other parts of the central nervous system.

Figure 2. A schematic cross-section showing the complex layering of the olfactory bulb into six strata: I, olfactory nerve layer; II, glomerular layer; III, external plexiform layer; IV, mitral cell layer; V, internal plexiform layer; VI, granular layer. GC, granule cell; MC, mitral cell; ON, olfactory nerve fibers from the olfactory receptors; OG, olfactory glomerulus; OV, olfactory ventricle (not universally present); PGC, periglomerular cells; PVZ, periventricular zone; TC, tufted cell [modified from Shepherd (1972)].

Thus, quantifying olfactory bulb volume involves combining different functional layers (and a central cavity, if grossly measured) into one measurement. Stephan et al. (1981) addressed this dilemma in two ways. First, the central cavity and its surrounding layer (called the periventricular zone) were eliminated from consideration since they do not contain neurons. Second, Baron et al. (1983) quantified olfactory bulb layers separately [also conducted in this manner by Bhatnagar et al. (1987) in humans]. Unfortunately, the layers are not easily separated and these investigators found that only certain combinations were morphologically quantifiable. The olfactory bulb volume data from Baron et al. (1983) are shown in a log-log plot against body mass [from Stephan et al. (1981)] in Figure 3. When all primate taxa are examined, there is a positive relationship between log volume of olfactory bulb layers [layers I and II combined; layer III; layers IV–VI combined, from Baron et al. (1983)] and log body mass (Fig. 3A). This positive relationship is also found if strepsirrhines (Fig. 3B) or anthropoids (Fig. 3C) are examined separately.

Figure 3. Olfactory bulb layers plotted against log body mass. Data plots based on Baron et al. (1983). A: The layers of the olfactory bulb plotted similarly for all primate taxa. The olfactory bulb layers plotted against log body mass are also shown for strepsirrhines (B) and anthropoids (C) separately. Tarsiers, which are of debated phylogenetic relationship to other primates, are shown in solid symbols of corresponding shape. In A, tarsiers were included in generating the regression lines, whereas in B and C, tarsiers were excluded. Note tarsiers fall more closely with anthropoids (C) than strepsirrhines (B). PVZ, periventricular zone; layers I and II, olfactory nerve layer and glomerular layer; layer III, external plexiform layer; layers IV to VI, mitral cell layer, internal plexiform layer, and granular layer.

From these plots, all layer combinations contribute relatively similarly to olfactory bulb volume (i.e., linear regression lines have similar slopes). However, the combinations of layers, dictated by morphologically identifiable boundaries, are rather unfortunate. One might expect that the olfactory axons of layer I have a volume that correlates well with olfactory epithelial volume (at least neuron body/axon counts would be identical). Unfortunately, the first two bulb layers are interwoven and impossible to distinguish for measurement as separate layers (Bhatnagar et al., 1987). Even if separable, the second layer contains elements of the secondary neural relays, and multiple axons from layer I can synapse at the same site within layer II. There are functional differences among layers IV–VI as well, although at least two of these three layers contain neuronal elements of secondary relays. Thus, overall olfactory bulb volume reflects more than a simple relationship to olfactory sensitivity. It is also a reflection of higher processing of olfactory stimuli; sensitivity and discrimination also relate to spatiotemporal context of bulb processing (Christensen and White, 2000).

The plots shown in Image 2 of Box 1 underscore that olfactory bulb volumes do increase in primates of larger body size. So, what is the basis of this relationship? First, we suggest that this may not reflect a causal relationship. It is arguable that a correlation may primarily exist between olfactory bulb volume and total brain volume; olfactory bulb volume may increase instead by virtue of increasing neural relays in brains of increasing size.

It is arguable that a correlation may primarily exist between olfactory bulb volume and total brain volume; olfactory bulb volume may increase instead by virtue of increasing neural relays in brains of increasing size.

Second, whether the size of neuronal bodies (e.g., mitral cells; Fig. 2) may be contributing to olfactory bulb volume differences between species has not been explored. As an additional factor, nonneuronal portions of the brain (i.e., ventricles) may differ markedly between taxa. For example, compare the periventricular zone volume in strepsirrhines (Fig. 3B) to anthropoids (Fig. 3C). In the former taxa, the periventricular zone is more positively related to body size; the volume of this zone increases with the size of the cavity it lines. Comparisons between anthropoids and strepsirrhines may be confounded by this difference, especially in regard to grossly acquired data or fossil studies that estimate olfactory bulb volume (Radinsky, 1979; Mirich et al., 2002; Takai et al., 2002).

CONCLUSIONS

This discussion leads us to conclude that the relevance of scaling to body size differs when measuring chemoreceptor structures compared with other neural tissues. It is very likely that body size corrections are unnecessary in many interspecific and intersex comparisons. These data require much more scrutiny, but even with the information we have at this juncture, we offer the following recommendations.

One, except for examination of growth or for comparisons among a broad range of mammalian taxa, interspecific data on olfactory elements should not be scaled to body size for comparisons. The relationship of the size of olfactory structures to overall brain size requires further exploration for this purpose.

Two, if neuronal body size and density are similar in chemosensory epithelia of species under study, absolute volumes of the epithelia may be compared (even if the body masses of the species are disparate). If neuron body sizes differ, comparisons of receptor neuron numbers are more appropriate. Ratios may be of value depending on the question (e.g., percentage of nasal epithelium that is chemosensory versus respiratory), but cranial rather than overall body dimensions are most meaningful for this purpose. Even in such cases, absolute values can be reported in addition to ratios.

Three, volume of the ventricles must be taken into consideration (i.e., should be subtracted) before inferring total olfactory bulb volume grossly or estimating it from fossils. Extant taxa may provide a guide during comparison of fossil taxa (comparisons within certain taxa may be largely affected or unaffected). A further complication arises since not all mammalian olfactory bulbs possess a ventricle.

Four, in defining macrosomia versus microsomia among different mammals, the vomeronasal system may deserve consideration for two reasons. First, the main and accessory olfactory systems are intimately tied during development and remain partially parallel in neural relays in adulthood. Second, it is increasingly apparent that at least some functional overlap of these two systems occurs (Dorries et al., 1997; Trinh and Storm, 2003). Thus, a possible definition of macrosomia may be that both olfactory and vomeronasal systems are present and functional. Given that some primates (e.g., lemurs and lorises) have functional VNO complexes while others (anthropoid apes, humans) do not, this distinction may be sufficient to separate those with the functional VNOs as macrosmatic and those without as microsmatic. Such a conclusion also relates to frugivorous and insectivorous bats, as presented by Bhatnagar and Kallen (1975).

In conclusion, large-scale taxonomic studies or growth studies benefit from body mass (or total brain volume) data for a comparative perspective. However, while such studies reveal important trends in the evolution of neural organization among species, they obscure certain issues. For example, relatively large visual centers and small olfactory centers in primates could reflect augmented vision, reduced olfaction, or both. Furthermore, our examination of proportional versus absolute measurements implies that in lower-level taxonomic comparisons, body size adjustments are needless and distort the data. This has implications for our understanding of olfactory abilities of primates. Although broadly speaking, primates may have reduced olfactory abilities compared to many other (macrosmatic) mammals, some specific generalizations require further scrutiny, e.g., that diurnal primates have reduced olfactory abilities compared to nocturnal primates. The question remains: is there a quantitative difference in olfactory abilities, or is this only an impression based on the relative importance of vision in the two groups?

Acknowledgements

The authors thank A. M. Burrows, M. I. Siegel, and M. A. Smith for invaluable comments on the manuscript during its development.