Plenty of Room at the Bottom?

Tiny solve problems of housing and maintaining oversized brains, shedding new light on nervous-system evolution

William G. Eberhard and William T. Wcislo

basic fact of life is that the size of shifts. The result is that different taxo- animals express the same kinds of be- A an ’s brain depends to some nomic groups have different, variant, havior as their large-bodied relatives. extent on its body size. A long history of versions of Haller’s Rule. Biologists have tended to ignore the studies of vertebrate animals has dem- The mechanisms that are responsible lower limits of body size and the physi- onstrated that the relationship between for grade shifts are only beginning to ological processes that are associated brain and body mass follows a power- be understood. But this combination with evolutionary decreases in brain law function. Smaller individuals have of generality and variability in Haller’s mass. Instead they have focused on evo- relatively larger brains for their body Rule appears to call into question some lutionary increases in brain size, and its sizes. This scaling relationship was pop- basic assumptions regarding the uni- possible links to intelligence and other ularized as Haller’s Rule by German formity of how the central nervous sys- mental processes. And almost all of the evolutionary biologist Bernhard Rensch tem functions among animals. It also current data have come from adults. But in 1948, in honor of Albrecht von Haller, reveals a number of overlooked design problems associated with the demands who first noticed the relationship nearly challenges faced by tiny organisms. Be- of a relatively large nervous system in a 250 years ago. Little has been known, cause neural tissue is metabolically ex- small animal are not limited to taxa with however, about relative brain size for pensive, minute animals must pay rela- miniaturized adults. Many species have invertebrates such as , spiders tively higher metabolic costs to power extremely small immature stages that and nematodes, even though they are their proportionally larger brains, and are free-living, and whose growth and among Earth’s more diverse and abun- they thus face different ecological chal- survival depends on their behavioral dant animal groups. But a recent wave lenges. There is reason to expect that capabilities. of studies of invertebrates confirms that tiny animals might cut corners wher- The new data on invertebrate brain Haller’s Rule applies to them as well, ever possible, for example by adopt- allometry have several important im- and that it extends to much smaller ing lifestyles that are behaviorally less plications. They challenge vertebrate- body sizes than previously thought. demanding. Yet available evidence in- based hypotheses that were proposed These tiny animals have been able dicates that at least some small-bodied to explain Haller’s Rule that invoked to substantially shift their allometric lines—that is, the relationship between their brain size and their overall body size—from those of vertebrates and other invertebrates. Animals that fol- low a given allometric line belong to the same grade and changes from one grade to another are known as grade

William Eberhard is professor of biology at the University of Costa Rica and a staff scientist at the Smithsonian Tropical Research Insitute (STRI) in Panamá. He received a Ph.D. in biology from Harvard University and has studied the biology of tropical for more than 40 years. William Wcislo is a staff scientist at STRI. He received his Ph.D. in entomology from the Uni- versity of Kansas and has conducted research on tropical insects for more than 25 years. Address for Eberhard: Smithsonian Tropical Research In- Figure 1. This Caenorhabditis elegans nematode, whose histone proteins and membranes are stitute, Apartado 0843, Balboa Panamá. E-mail: stained here, has a nervous system that is a marvel of miniaturization. The 1-millimeter-long [email protected] worm contains only 302 neurons. Yet with this neural network it executes a surprisingly wide

226 American Scientist, Volume 100 factors such as surface-volume relations, mites, the relatively large brain takes up Another approach to solving the longevity and metabolic rates. They also so much room that it overflows into the housing problem would be to reduce challenge the idea, again derived from legs, giving new meaning to the phrase neuron size, thus reducing overall brain studies of vertebrates, that animals with “thinking on your feet.” size while maintaining similar numbers relatively and absolutely larger brains In some groups of tiny insects, such of neurons and the degree of connectiv- have more sophisticated behavioral as strepsipterans, the shape of the brain ity. The sparse available data suggest, abilities and mental capabilities. For is modified to pack it tightly against in- however, that such adjustments are these reasons, the time is ripe for explor- ternal structures and muscles. Although incomplete, and that within any given ing the ways that central nervous sys- the “brain” is conventionally construed taxon, smaller animals usually have tems are organized at very small scales, to be that part of the central nervous fewer neurons. Reductions in neuron within constraints that differ from those system that is housed in the head, some size are possible, but only to a point. No- of large-bodied organisms. tiny and other insects blur this bel laureate physicist Richard Feynman distinction because they displace some discussed general limitations regarding Problems of Miniaturization or all of their large brains from the head storing and retrieving information at By focusing on evolutionary increases to the thorax or even to the abdomen. In extremely small scales and concluded in brain size, biologists have general- general the anatomical trade-offs that that “there’s plenty of room at the bot- ly overlooked a basic miniaturization result from the displacement of other tom” when building artificial informa- problem that follows from Haller’s Rule: tissues have not been identified, nor tion-processing systems at nano­scales. Where can a relatively large brain fit in have their costs been determined. The In biology, however, neuron-based in- a small body? In salamanders and fish, design changes imply that some fea- formation-processing systems bottom for example, the brain is housed within tures are sometimes sacrificed to house out at sizes where Feynman just gets go- a cavity formed by skull bones. In min- enlarged central nervous systems in ing. There is a theoretical lower physi- iaturized forms, some of these bones are very small animals, which may play a cal limit on the functional diameter of lost or reduced in size, freeing up room role in setting the lower limits of body an axon (about 0.1 micron). Below this for the relatively large brain. Arthropods size in a given taxon. A minute hooded size, an axon can no longer transmit reli- have external skeletons, which they can (; ), for able information because the signal is deform to some extent to create more example, has fewer muscles in its head swamped by noise from spontaneous internal space. A nymph of the orb- and thorax than do larger related bee- depolarizations of the membranes. In weaving spider, Anapisona simoni, with a tles. The space taken up by the enlarged addition, the minimum size of a neu- body mass of less than 0.005 milligram, brain of a nymph of the jumping spider ron cell body is limited by the size of its appears as a speck of dust to the un- Phidippus clarus comes at the cost of re- nucleus, which in turn is limited by the aided eye. In these minute orb weavers, duced space for digestive diverticula. animal’s genome size. The nucleus com- nearly 80 percent of the cephalothorax In both cases it is likely that there are prises up to 80 to 90 percent of the vol- is filled with the brain. To house their as yet undetermined costs associated umes of small neuron cell bodies in tiny relatively large brains, minute spiders with these changes. If, for instance, it insects. One route to miniaturization (including tiny spiderlings of species were advantageous for an adult spider among arthropods would be to delete with large-bodied adults) have a con- to have digestive tissue in the cepha- chromosomes, pack the chromatin more spicuous outward bulge in the sternum, lothorax, then a similar design would tightly or eliminate the nucleus, which which increases the internal volume of seem likely to have been advantageous would permit a smaller neuron. Such the cephalothorax, where brain tissue is for a nymph if it were not encumbered modifications had been documented in housed. In some species of spiders and with a relatively large brain. vertebrates but were unknown in inver-

range of behaviors, including multiple types of learning. The contrast between this and the neuron-profligate nervous systems of animals such as vertebrates suggests that the dynamics of nervous system function can differ radically. In this article the authors explore how such variability may arise and its implications. (Image courtesy of Ian D. Chin-Sang of Queen’s University, Kingston, Ontario.) www.americanscientist.org 2012 May–June 227 4

2

0 salamanders other amphibians –2 mammals spiders ants

log brain –4 bees mites kinorhynch –6 various insects beetle larva –8 nematode copepod nauplius weevils –10 –8 –6 –4 –2 0 2 4 6 8 log body

4 –1. 0

2 –1. 5

0 salamanders –2.0 other amphibians –2 mammals honey bee spiders –2.5 ants

log brain –4 bees

log brain/body –3.0 mites kinorhynch –6 various insects –3.5 beetle larva –8 nematode –4.0 copepod nauplius weevils –10 –4.5 –8 –6 –4 –2 0 2 4 6 8 –8 –6 –4 –2 0 2 4 6 8 log body log body

Figure 2. Spanning nearly 15 orders of magnitude, these plots provide the most comprehensive comparisons available of brain mass versus body–1. mass 0 for animals (left), and the fraction of total body mass devoted to the brain versus body mass (right), expressed as logarithms. The second plot shows the extent of grade shifts between taxa. (Vertebrate data are from Georg Striedter’s 2005 Principles of Brain Evolution, and invertebrate data from various sources cited in the authors’ 2011 article in Advances in Physiology.) –1. 5 tebrates until just recently. Alexy Polilov compensatory mechanisms that en- tions as a larger brain and is thus more of –2.0Lomonosov Moscow State University able this behavior with only a greatly costly on a unit basis. These differences has shown thathoney most bee of the neurons of reduced number of neurons, most of imply that the strength of natural se- minute–2.5 parasitic wasps, Megaphragma which lack nuclei, are not known. lection for energetic efficiency may be sp. (), with body more intense in small than in large spe-

lengthslog brain/body –3.0 of 170 to 200 micrometers, lack The Price of Brain Upkeep cies, especially within a given grade. nuclei. The pupal central nervous sys- Energetically, the brain is a gas-guzzler, They also raise the question of how tiny tem–3.5 has about 7,400 nuclei, but near the as neural tissue is more expensive to animals pay these high energetic costs. end of pupal development most neuro- run than most other tissues. Humans, With respect to information process- nal– 4.0cell bodies break open, or lyse, and for example, have brains that account ing, animals might adopt several non- lose their nuclei. The adult central ner- for slightly more than 2 percent of our exclusive strategies to reduce costs. vous–4.5 system, therefore, has about 7,000 biomass, but they burn up more than 15 One, the size limitation option, is to re- cells without–8 – nuclei,6 –4 and– only2 3390 to 3722 percent4 6 of our8 basal metabolic energy. duce behavioral capacities and thus re- cells with nuclei, of which onlylog body 179 to In addition, the density of information duce the amount of neural tissue need- 253 are in the brain. processing is greater in a smaller brain ed to sustain behavior. Another is the Lysis also is associated with volu- when it is performing the same opera- oversized brain option, which involves metric changes in the nervous system. The pupal brain volume of about 93,600 cubic micrometers decreases to 52,200 cubic micrometers in the adult. In ad- dition, numerous folds are present in the cuticle of the back of the head, the occipital area, and the size of the head capsule in this area is reduced. Remark- ably, the central nervous system of M. mymaripenne has orders of magnitude fewer neurons in comparison with oth- er flying insects, such as Musca flies, which have 340,000 neurons. A some- 6 millimeters what larger trichogrammatid wasp, 6 millimeters Trichogramma evanescens, for example, Anapisona simoni has 37,000 neurons in just one part of the brain, the supraesophageal ganglia. Figure 3. TheseAnapisona webs were made simoni by two orb-weaving spiders, Allocyclosa (left) and Despite their extreme central nervous Anapisona (above) that differ greatly in body system modifications, Megaphragma size. Detailed measurements of the webs wasps nevertheless perform behavior show that extremely small spiders, with tiny such as mating, flying, host searching brains, execute complex tasks of web build- and recognition, although the details ing with no less precision than larger spiders. 6 millimeters have not been studied. The kinds of (Photographs courtesy of the authors.) 6 millimeters 228 American Scientist, Volume 100 Allocyclosa bifurca Allocyclosa bifurca ovarioles maintaining behavioral capacities and heart bearing the high metabolic costs of an enlarged central nervous system. And then there is the economy-of-design cerebrum digestive tract option, or modifying the properties of neurons and neural networks to in- crease behavioral capability per unit of nervous tissue. The last option could allow smaller animals to produce com- parable behavior without investing in large nervous systems. There are several ways that animals might achieve economy of design and thus economize on the energy budgets of their brains. All animals have sen- sory receptors that are tuned to spe- cific inputs to improve their efficien- suboesophageal and meso- and meta- cy. Matched filters take advantage of prothoracic ganglia thoracic ganglia mechanical properties of sensors and Figure 4. Anatomical trade-offs must be made to accommodate relatively large central nervous stimuli to filter sensory input without systems within the relatively small bodies of invertebrate animals. This reconstruction of an expending energy to do so, which re- adult hooded beetle, , shows that the brain was shifted almost entirely duces processing expenses at the recep- from the head to the prothorax, displacing other structures. The central nervous system is light tor and at higher central nervous sys- blue, the exoskeleton is rust, muscles are brown, the digestive system is yellow, the heart is red tem levels. Many insects, for example, and the reproductive system is violet. (Image adapted from Alexy Polilov and Rolf Beutel’s are highly sensitive to polarized light, 2010 article in Structure and Development.) a trait associated with the physical alignment of rhodopsin molecules in elegans has been analyzed with respect ioral subroutines can facilitate goal-di- the microvillar membranes of photo- to an optimal design in this save-wire rected behavior if subroutines are seri- receptors. There are numerous other aspect. The trend toward greater fu- ally ordered to maximize desired motor possible opportunities to reduce central sion of different ganglia in the central output per bit of sensory input. For in- nervous system costs, such as a heavier nervous system of minute insects may stance, grasping a randomly positioned reliance on analog transmission and also be related to this type of efficiency. object is a difficult task that requires a graded depolarizations, which work The relative frequencies of such design series of visual-to-motor transforma- at small distances and are energetically features in large versus small animals tions. In robots this task is rendered eas- more efficient than the action potentials are not known. ier computationally if a robot executes that many other animals use for com- Other possibly important differences a subroutine to turn and face the object, munication among neurons. Another in small nervous systems involve the triggering a second subroutine to move example includes the use of the same neurons themselves. In at least some to a specified distance from it. Grasping neurons for multiple tasks, such as both groups, the dendrites of smaller neu- it from a standard distance and direc- sensory and motor functions, as is com- rons are themselves smaller and said tion simplifies the task. The possibility mon in nematodes. Neuromodulation to be less complex. Nearly 100 years that small animals are more likely to may be used more frequently by small ago, insect neuroanatomist Santiago use such behavioral designs has, to our animals. That involves altering the ef- Ramón y Cajal produced pioneering— knowledge, never been checked. fects of a neuronal circuit by exposing and beautiful—studies of the nervous it to different chemical environments, systems of insects. He found insect neu- Grade-Change Mysteries producing different behaviors. Nema- rons to be more elaborate than those Grade changes reflect increases or de- todes frequently use muscle plates that of vertebrates, and likened the neuro- creases in the slopes and elevations allow a single synaptic process to stim- anatomy of an insect to a “fine pock- of brain-body scaling relationships ulate multiple muscles. Still other strat- et watch,” as opposed to the “rough in different taxa. Evolutionary transi- egies include the indirect control of cilia grandfather clock” neuroanatomy of tions between grades are poorly un- through muscles; a reduction in the rel- a vertebrate. In general, the functional derstood, but are obviously crucial to ative numbers of interneurons, which significance of these neuroanatomical understanding how nervous systems transmit signals from one neuron to differences is not well understood, but function and how the diverse array of another as opposed to sensory and mo- we expect that they represent promis- animal body sizes has evolved. Most tor neurons; and the rearrangement of ing starting points for future research. previous discussions of grade changes neuron positions and connections in Design economies could also occur have emphasized evolutionary transi- order to minimize the total length of at the behavioral level but are even less tions toward larger body size and more axons and dendrites. That last design studied. Robotic engineers incorporate highly encephalized forms, probably is analogous to an architect holding behavioral design features to minimize due to a general fascination with the down building costs by minimizing the the input needed to generate informa- possible implications for greater intel- length of wire needed to provide elec- tion about the external world in order ligence. The new invertebrate data high- tricity throughout a house. The nervous to achieve desired outcomes. A rigid light evolutionary changes in the op- system of the nematode Caenorhabditis hierarchy of relatively simple behav- posite direction—miniaturization—and www.americanscientist.org 2012 May–June 229 populations of neurons—on the activity patterns in recurrently interconnected populations of neurons—rather than on the activity of individual cells. The activity of individual neurons is nei- ther consistent nor useful in processing information, and patterned activity of neuronal populations can emerge in al- ternative ways from different subsets of the population. A single dysfunctional neuron can thus be inconsequential for a vertebrate but catastrophic for a nem- atode. The loss of one particular neuron in C. elegans, for instance, brings evolu- tionary fitness to zero because it leaves females unable to lay eggs. The striking visual resemblance be- tween the circuits on a computer chip and a wiring diagram of nerve fibers in the ventral nerve cord of a nematode draws attention to the computer-like traits of a nematode’s nervous system, with its fixed number of neurons with invariant connections. This design central nervous system stands in contrast to the decidedly non- computer-like traits of vertebrate ner- digestive diverticuli vous systems. The processes by which these two extreme types of nervous sys- Figure 5. Anatomical trade-offs sometimes must also be made by immature invertebrates tems are built up as an organism grows when compared to adults of the same species. Sagittal sections of the cephalothorax in each are also strikingly different. Growth of of three developmental stages of the jumping spider Phidippus clarus (Salticidae) illustrate the vertebrate nervous system is char- changes to body designs made necessary early on by a relatively larger central nervous system acterized by an extraordinary overpro- (blue). As is typical in other spiders, the central nervous system occupies a larger fraction of duction of neurons in young stages, the cephalothorax volume in smaller individuals. In this salticid, digestive tissue (red) that is followed by selective pruning of inac- abundant in the adult cephalothorax is nearly completely missing from the cephalothorax in the second instar, the first developmental phase that lives outside the egg sac. (Image adapted tive neurons and synapses. The extent from David E. Hill’s 1975 master’s thesis at Oregon State University.) to which neurons die during develop- ment varies among, and within, spe- cies. For example, the percentage of the attendant central nervous system by comparing two groups that are ex- retinal ganglion cells that die during design problems at minute body sizes. tremely different—the neuron-miserly development varies from 80 percent in A grade change can make it possible nematodes, and the neuron-profligate cats to 60 to 70 percent in rats, mice, for much smaller body sizes to evolve vertebrates. The nematode C. elegans rhesus monkeys and humans, and to than would have been possible if the has a nervous system with only 302 approximately 40 percent in chickens animals had continued to use the slopes neurons, and some other nematodes and amphibians. Selective pruning of and elevations of brain-body lines from and tiny invertebrates have even lower neurons in nematodes is, on the other the previous grades. For example, if a numbers. Each C. elegans neuron has hand, nearly nonexistent. In C. elegans a 1-milligram animal used the scaling rule only about 25 synapses, and the neu- total of 8 of 310 neurons, 2.6 percent, are used by salamanders, it would have a rons are connected in highly consistent discarded as the animal matures, and brain that constituted a prohibitive 20 and relatively simple ways. they are always exactly the same cells, percent of its body, a proportion that is The contrast between the brain of a resulting in the 302 neurons of the adult larger than that of any known animal. nematode and that of a human could hermaphrodite; there is no indication Grade shifts suggest that some partic- hardly be greater. Our brains have as- that use or disuse is a factor influencing ular taxa of small-sized animals, such tronomical numbers of neurons, an es- cell death. Similarly, a second instar spi- as ants, have apparently solved scaling timated 85,000,000,000; huge numbers derling and an adult of the orb weaver problems that seemed insuperable for of synapses per neuron, such as 10,000 Argiope aurantia have approximately the animals of larger-sized taxa, although per pyramidal cell in the cortex; and same number of neuronal cells, despite they do not explain how or why. extremely high degrees of connectivity. an approximately 24-fold difference in The changes in design that are asso- For example, there are roughly 101,000,000 total brain volume. Such extreme con- ciated with most grade shifts remain to possible circuits in the human cortex trasts raise the possibility that nervous be worked out, but it is likely that they alone, which prompted Nobel laureate system function differs profoundly in are at least sometimes associated with Gerald Edelman to describe the human different parts of the animal kingdom, the evolution of new neural design brain as “the jungle in the head.” The in contrast to the generality of much of mechanisms. This can be illustrated functioning of our brain depends on biochemistry, molecular genetics and

230 American Scientist, Volume 100 molecular development. The miserly design typified by nematodes may rep- resent adaptations that permitted the evolution of miniature body sizes.

Haller’s Rule, Grades, Behavior The common supposition derived from vertebrates—that animals in lower grades are necessarily inferior or less capable in their behavior—is not well supported by facts. Among adult orb- weaving spiders that vary in body mass by 400,000 times, including those near the lower size limit for spiders, there is no evidence of inferior performance with respect to the precision of execut- ing certain web construction behaviors or the ability to adjust web designs to local conditions or to build alternative web designs. Qualitative observations also suggest that size limitation of be- havior fails to explain some other grade differences, such as that between honey bees and salamanders. Miniature sala- manders and honey bees have approxi- mately the same body size, but the lat- ter fall on an allometric line lower than that of the salamander. A honey bee is nevertheless capable of such feats as navigating using landmarks and polar- ized sunlight that is adjusted for time of day; learning complex patterns in- volving different sensory modalities or general concepts such as different or similar and above or below; and us- ing language to tell nest mates where to find food. It would be difficult to defend the argument that a honey bee is behaviorally inferior to a miniature salamander. Even the neuroanatomically simple nematodes behave in ways that are not fundamentally different from many animals with many more neurons. For Figure 6. Nobel Prize–winning insect neuroanatomist Santiago Ramón y Cajal produced pio- example, C. elegans senses and responds neering and beautiful studies of insect nervous systems. This elegant 1915 schematic is a dia- to various stimuli, including physical gram of the complex anatomy of neurons in a honey bee’s retina and optic lobe. (Image cour- contact with environmental objects, tesy of Trabajos del Laboratorio de Investigaciones Biológicos de la Universidad de Madrid.) and perceives various chemicals, oxy- gen concentration, osmolarity, pH, tem- tion, nematodes can learn to modify a do we quantify and compare behavior perature, light and pheromones. These variety of motor behaviors on the basis in biologically meaningful ways across various inputs are used to coordinate of their experience. different taxa? The size of an animal’s motor output, and to evaluate condi- behavioral repertoire has sometimes tions such as the density and sex of con- Problems Measuring Behavior been used as a quantitative measure of specifics. Motor outputs include dif- The idea that grade changes are asso- a species’ behavioral complexity. This ferent movements for swimming and ciated with differences in behavioral intuitively appealing idea suffers from for crawling on a surface, for turning capabilities in different species stems several problems. In a pioneering study or reversing those movements or for from an assumed correlation between relating behavior to head and brain size altering them after fixed time periods. brain size and behavioral capabilities. in ants, Blaine Cole, now at the Univer- These animals orient and move toward Comparative studies have often been sity of Houston, highlighted some of or away from point stimuli; forage for confounded by the use of imprecise, them. Such a metric relies on subjective food, engulf it, perform rhythmic swal- difficult-to-define behavioral metrics decisions to distinguish specific behav- lowing movements and defecate; seek such as intelligence. And an important iors, and faces a problem common in mates, copulate and lay eggs. In addi- question remains unanswered: How artificial classification systems. Observ- www.americanscientist.org 2012 May–June 231 and hence less reliable information about the environment. They have less thorough processing of sensory inputs because of fewer interneurons or den- drites. Motor output or coordination among different limbs may be compro- mised by reduced feedback from pro- prioceptors, which sense stimuli from within a body, or from increased noise in the nervous system. The available data from tiny orb weaving spiders do not suggest behavioral limitations and iStockphoto are not consistent with the size-limita- Figure 7. The diagram on the left shows the connections of 70 of the 134 fibers in the circum­ tion hypothesis. Tiny spiders are mor- enteric ring near the origin of the ventral nerve cord of the nematode Ascaris megalocephala. It looks a lot like the architecture of a modern computer chip. The uncanny resemblance phologically modified to house a rela- emphasizes the computer-like traits of nematode nervous systems that include a fixed num- tively enlarged brain, suggesting that ber of elements and unchanging connections between them. A human brain, in contrast, is they have adopted strategies consistent emphatically not computer-like. The number of neurons and their connectivity vary dramati- with the over-sized brain alternative we cally. The nematode-human comparison suggest that there may be profound functional and have described. The generality of this organizational differences in the brains of taxa in different grades. (Diagram on the left from finding is currently unknown, how- Advances in Insect Physiology, Volume 40, Elsevier). ever, because of the lack of compara- tive data. Other behaviors have not yet ers who are “splitters” will recognize a cision in adjusting behavior to other been thoroughly explored in spiders larger number of behaviors than those variables. These traits can be compared and other tiny animals. One that might who are “lumpers.” In addition, reper- more consistently across diverse spe- be meaningfully compared among dif- toire comparisons rely on a number of cies. Behavioral precision, the ability ferent species is the ability to learn and unverified assumptions: that each be- to accurately and consistently repro- remember different types of lessons. havior is equally demanding in terms duce the same behavior, has been hy- Finally, the use of overly inclusive of neural system processing, that behav- pothesized to be less developed in rela- measurements, such as overall brain iors called the same thing in different tively small-brained animals. Mistakes size, rather than measurements of re- species are equally demanding and that or imprecision might arise in smaller gions involved directly in behavioral rare behaviors, which are more likely to animals in several ways. For one, they tasks, can confound comparisons among be missed by observers, are not drivers have fewer sensory receptors, and thus species. Bats, for example, rely on hear- of brain evolution. It’s also assumed that should have decreased sensory input ing to a greater extent than we do. Due the speed and precision with which a given behavior is performed is the same in different species and that the environ- mental influences shaping behavioral expression are minimal. It also makes the unlikely supposition that laboratory colonies will reveal the full repertoire of important behaviors. These problems have led some re- searchers to adopt other metrics, such as the frequency of mistakes when making decisions, or the degree of pre-

Figure 8. A honey bee has a brain with ap- proximately 850,000 neurons, yet a bee is ca- pable of sophisticated behavior, rivaling or surpassing that of many vertebrates. Here a bee communicates a highly abstract message by performing a figure-eight waggle dance,

using the angle between the straight run of Kim Taylor/Warren Photographic Ltd the dance and the direction of gravity as a ref- erence to convey the location of food relative to the position of the sun. The duration of the straight run conveys distance from the hive. This image was made with high-speed strobe equipment that produces a burst of nine flash- es within 20 milliseconds. As a result, the rap- idly moving messenger bee (center) appears in a series of overlapping images while those observing appear to be more or less still.

232 American Scientist, Volume 100 to body size differences, our subcorti- known adult insects, perhaps because manders to beetles, so consistently have cal auditory brain region is vastly larger their larvae hatch into a host environ- relatively larger brains when they have than that of bats, so overall size is not ment that contains all the necessary nu- smaller bodies? The new invertebrate informative. Relative size, however, re- trients, thus permitting greatly reduced data help by ruling out some previous veals that the subcortical auditory region egg sizes and lower energy reserves. explanations that were only reasonable of the bat brain comprises 1.6 percent of Small immature stages of species may for particular groups of vertebrates. But total brain volume, versus 0.015 percent represent the leading edge of evolution- we do not have an alternative general in humans. The problem of associating ary innovations to solve size-related explanation. It is very unusual in biolo- behavior with particular brain regions physiological problems as a lineage gy that such a general trend as Haller’s is exacerbated by cultural differences evolves toward smaller body size. Are Rule should have such a depauperate among scientific disciplines. Studies grade changes associated with innova- array of hypotheses lined up as possible on behavior often take a comparative tions that result in energetic efficiencies explanations. More often the problem is approach and include data from di- in the central nervous system? For in- having too many competing ideas to verse species. Neurobiological studies stance, weevils have a low allometric test. Understanding why Haller’s Rule are more often focused on a very small brain-body line compared with many is so generally true is likely to be im- number of so-called model organisms other insects. Are there economies of de- portant in answering central questions that are studied under laboratory condi- sign that make their nervous systems regarding the evolution of the central tions, where they express a much small- more efficient in generating behavioral nervous system and how and why dif- er range of behaviors than in nature. abilities? And are these efficiencies asso- ferent grade changes evolved. Indeed, one recent study showed that ciated with the evolutionary and ecolog- approximately 75 percent of the research ical success of weevils, one of the most Bibliography efforts of neuroscientists, as judged by speciose taxa of all animals? Answers Bonner, J. 2006. Why Size Matters. Princeton, numbers of publications, were directed to such questions are unknown, in part NJ: Princeton University Press. at only three species—the mouse, the because they are rarely asked. We agree Eberhard, W. G. 2011. Are smaller animals be- haviorally limited? Lack of clear constraints rats, and the human. When compared with Princeton biologist John Bonner, in miniature spiders. Animal Behaviour with a recent estimate of 7.7 million ani- who emphasizes in a recent book that 81:813–823. mal species worldwide, that is about 3.9 “size matters” in ecology and evolu- Eberhard, W. G., and W. T. Wcislo. 2011. Grade x 10-5 percent of animal biodiversity. tion. An understanding of the evolu- changes in brain-body allometry: Morpho- tion of brain form and function requires logical and behavioral correlates of brain size in miniature spiders, insects and other Significant Consequences comprehensive data on neuroanatomy, invertebrates. Advances in Insect Physiology Problems associated with miniaturiza- neurophysiology, behavior and ecology 40:155–214. tion are much more general than they from a diverse array of species, not just Grebennikov, V. V., and R. G. Beutel. 2002. might first appear. In addition to the a few models. Studies dealing with the Morphology of the minute larva of Ptinella many species with miniaturized adults, nervous systems and behavior of very tenella, with specific reference to effects of there are many more species with small animals are likely to reveal phe- miniaturization and the systematic posi- tion of Ptiliidae (Coleoptera: Staphylinoi- moderate-sized adults that have very nomena not seen in the larger animals dea). Arthropod Structure and Development small, free-living, immature stages. A that are typically studied. Tremendous 31:157–172. mature female of the giant golden-orb- opportunities await for neuroethologi- Hanken, J., and D. B. Wake. 1993. Miniaturiza- weaver Nephila clavipes weighs on the cal studies of taxa with miniature ani- tion of body size: Organismal consequences order of 2,000 milligrams, for example, mals, and syntheses of data and ideas and evolutionary significance. Annual Re- view of Ecology and Systematics 24:501–519. but each of her newly emerged spid- from disparate fields such as brain al- Misunami, M., F. Yokohari and M. Takahata. erlings weighs only 0.7 milligram and lometry, animal behavior, ecology, neu- 2004. Further exploration into the adaptive has tell-tale adjustments to small size, robiology, classic invertebrate zoology design of the arthropod ‘‘microbrain’’: I. including extensions of its brain into a and molecular and developmental biol- Sensory and memory-processing systems. bulging sternum and into its legs. The ogy. An understanding of the patterns Zoological Sciences 21:1141–1151. Polilov, A. A., and R. G. Beutel. 2009. Minia- ecological problems relating to ener- and processes involved in evolutionary turization effects in larvae and adults of gy intake and expenditure of smaller, decreases in body and brain size is also Mikado sp. (Coleoptera: Ptiliidae), one of young individuals are likely to be dif- likely to illuminate those associated with the smallest free-living insects. Arthropod ferent from those of larger conspecific evolutionary size expansion, and both Structure and Development 38:247­–270. adults. Because of their higher meta- are likely to further our understanding Quesada, R., E. Triana, G. Vargas, M. Seid, J. Douglass, W. G. Eberhard and W. T. Wcislo. bolic costs, smaller animals may be of crucial evolutionary transitions from 2011. The allometry of CNS size and conse- more likely to be living on the edge of one evolutionary grade to another. quences of miniaturization in orb-weaving energetic limitations and less buffered We can’t conclude this discussion of spiders. Arthropod Structure & Development against temporary food shortages. An brain and body sizes without acknowl- 40:521–529. increased susceptibility to unfavorable edging that we have ignored the pro- energy balances could have significant verbial 500-pound gorilla lurking in the biological repercussions, such as limit- room throughout this discussion. We ing an animal’s geographic distribution have examined many different conse- For relevant Web links, consult this or its ability to survive food shortages quences of Haller’s Rule, but we have ­issue of American Scientist Online: or other stresses, and might select for given no explanation for why the rule http://www.americanscientist.org/ alternative ecological strategies. Parasit- should be true. Why should organisms issues/id.96/past.aspx ic wasps () that use insect ranging from the different castes of ants eggs as hosts are among the smallest in a single nest, from primates to sala- www.americanscientist.org 2012 May–June 233