Leaf Mottling: Relation to Growth Form and Leaf Phenology and Possible Role as Camouflage Author(s): T. J. Givnish Source: Functional Ecology, Vol. 4, No. 4 (1990), pp. 463-474 Published by: British Ecological Society Stable URL: http://www.jstor.org/stable/2389314 Accessed: 27-04-2016 13:23 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

British Ecological Society, Wiley are collaborating with JSTOR to digitize, preserve and extend access to Functional Ecology

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms Functional ESSAY REVIEW Ecology 1990, 4, 463-'474 Leaf mottling: relation to growth form and leaf phenology and possible role as camouflage

T. J. GIVNISH tion or masking of chlorophyll over portions of the Department of Botany, University of Wisconsin, upper leaf surface, often in a more or less mottled Madison, Wisconsin, USA pattern. Leaves mottled in various shades of green, or marked by splashes or stripes of white (or sometimes purple, red or pink) are characteristic Abstract. The arguments of Smith (1986) bearing of many species of angiosperms, particularly on the adaptive significance of leaf mottling are among understorey herbs in temperate and tropi- criticized and an alternative hypothesis is pre- cal forests (Neger, 1913; Richards, 1952; Morgan, sented. I propose that mottling may serve to 1971; Smith, 1986; Givnish, 1988a; Evenari, 1989). camouflage the foliage of certain groups of short- Smith (1986) discusses some of the selection statured forest herbs, by disrupting their outline as pressures that may favour mottled leaves, in terms perceived by colour-blind vertebrate herbivores in of processes that could help maintain a population sun-dappled understoreys. Certain phenological polymorphic for mottled and unmottled leaves. He groups are likely to be particularly vulnerable to hypothesizes that mottled leaves might be advan- herbivores, based on their high leaf nitrogen con- tageous by: (i) serving to mimic leaf miner damage tent (spring ephemerals, spring leaves of summer- and thereby deterring actual oviposition by active species), leaf activity when few other miners, which are posited to avoid the reduced species possess foliage (evergreen species, resource levels or induced defences (Schultz & wintergreen species, winter leaves of dimorphic Baldwin, 1982) that may be associated with mined species) and/or relative cost of replacing con- leaves; (ii) reducing leaf temperature and transpir- sumed foliage (evergreen species on sterile soils). ation in sunny, hot or dry microsites; (iii) visually These groups are also exposed to relatively high diverging from other foliage sought by visually irradiances and so are less likely to suffer photo- orienting herbivores (see Gilbert, 1975 and synthetic losses as a result of the reduced leaf Rausher, 1981 for similar suggestions based on absorptance that accompanies mottling. A survey differences in leaf shape); and/or (iv) serving as of the incidence of leaf mottling in the native flora aposomatic coloration, advertising a chemical, of the north-eastern USA supports these ideas: mechanical or biological defence. Set against these mottled leaves occur almost exclusively among potential advantages is (v) the reduction of light forest herbs and are substantially over-represented interception likely to result from the masking or among evergreen, wintergreen, and spring ephem- absence of chlorophyll over the mottled areas of eral species, and among the winter leaves of the leaf. dimorphic species and the spring leaves of Smith (1986) presents data that bear on these summer-active species. Mottled leaves are es- hypotheses for Byttneria aculeata Jacq. (Stercu- pecially common among evergreen forest herbs. liaceae), a weedy tropical subcanopy vine growing Testable extensions of the camouflage hypothesis in Panama. Juvenile plants of Byttneria have either to account for the relative abundance of mottled mottled or unmottled leaves. Mottled plants are leaves among tropical forest herbs are discussed relatively more frequent in partial or complete and parallels with other recently proposed clearings than under closed canopies and trans- mechanisms of vegetative crypsis in angiosperms plant studies suggest that the degree of leaf mot- based on leaf size, shape, texture, colour, and/or tling is genetically determined. Leaf production by movement (i.e. Mimosa pudica) are outlined. transplants in the absence of herbivory indicates Key-words: Anti-herbivore defence, camouflage, forest herbs, leaf mottling, leaf phenology that the average amount of variegation per leaf is correlated positively with growth in open habitats and negatively with growth in shaded sites; Smith Introduction interprets this as tentative support for hypotheses 463 Leaf variegation entails the local absence, reduc- (ii) and (v). Furthermore, when open and shaded

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 464 sites are pooled, the mottled morph is less heavily preference for shady conditions or for plants T. J. Givnish attacked than the unmottled morph when they grown under shady conditions. To exclude this occur at the same frequency (the extent of leaf possibility, we would need studies of miner attack mining on a given morph increases with its rela- as a function of leaf variegation in sunny and tive frequency). Smith (1986) interprets the latter shady field sites in which the frequencies of finding as tentative support for hypothesis (i). mottled and unmottled vines have been experi- mentally manipulated. 3 Finally, Smith's (1986) proposal that the Critique of Smiths's model for leaf combined effects of leaf variegation on leaf miners, variegation light interception and heat load favour mottled Smith's (1986) empirical findings are important leaves in open or partially shaded sites contradicts and appear to provide the first experimental evid- qualitative observations by Richards (1952) and ence for the adaptive significance of leaf varie- Givnish (1988a) that species characterized by such gation in sunny vs shady environments. Yet there leaves are relatively most numerous among herbs are some difficulties with the interpretations he native to temperate and tropical forest under- advances: storeys. Smith's proposal may thus have no gen- 1 The growth of transplanted Byttneria vines is in eral applicability and the correlation between leaf accord with hypotheses (ii) and (v) that mottled mottling and open or partially shaded habitats (and hence more reflective) leaves would be cooler seen in Byttneria may belie the pattern seen across and achieve lower rates of transpiration in sunny plant species elsewhere. sites, but achieve substantially lower rates of photosynthesis in shady sites. Yet other plausible Camouflage hypothesis interpretations of this result exist. For example, a As an alternative hypothesis, I propose that mot- correlation between the degree of leaf variegation tling may serve to camouflage the foliage of certain and growth rate in forest openings may arise if groups of forest herbs, by disrupting their outline variegation in Byttneria is genetically linked to as perceived by colour-blind vertebrate herbivores other, functionally unrelated traits (e.g. Rubisco (e.g. deer) in sun-dappled understoreys. It is expli- level, stomatal conductance, leaf thickness, leaf- citly assumed that, when viewed in black-and- root allocation, anti-herbivore defences) that white, mottling acts as a dappled camouflage determine photosynthetic performance and/or pattern against the sunlight-dappled forest floor competitive ability in sunny sites (see Bj6rkman, and would tend to reduce the probability that a 1981; Givnish, 1988a). mottled leaf or spatial cluster of such leaves would 2 More importantly, even if hypotheses (ii) and (v) be consumed. Such crypsis is possible only for were both correct, neither would support an adap- leaves held close to the ground, because even tive value for leaf mottling per se, because leaves of foliage that matches the forest floor light pattern uniformly reduced absorptance can reflect or perfectly could be detected visually if it were held transmit as much light, on average, as mottled much above the ground, through differences in the leaves. The same is true of hypothesis (iv): it is not focal planes of leaves vs ground relative to a clear why aposomatic leaves should be mottled, stationary herbivore or differences in parallax rather than strikingly marked in bold patterns. No relative to a moving herbivore. evidence is presented to support hypothesis (iii) If this hypothesis is correct, then leaf mottling and it is not clear why visual divergence through should be most common in forest herbs. An leaf mottling (vis-&-vis leaf size or shape) should important corollary of this hypothesis, indepen- be particularly common in shady environments. dent of previous observations, would be that Of Smith's hypotheses, only (i), involving among forest herbs leaf mottling should be most supposed mimicry of leaf miner damage, entails an common in groups in which plants are exposed to advantage specific to leaf mottling. Data pooled the greatest threat of herbivory, in which the across transects do indicate a generally lower camouflage effect would be greatest, and/or in incidence of mining in mottled Byttneria leaves. which the negative impact of leaf variegation on Yet, as Smith (1986) himself notes, this lower photosynthesis would be least. incidence in mottled leaves might simply reflect a Certain phenological groups are likely to be tendency by the Byttneria miner to attack the particularly vulnerable to herbivores, based on locally more common leaf morph preferentially, their high leaf nitrogen content (e.g. spring ephem- given that the most common morph in the shade is erals, spring leaves of summer-active species) or unmottled and provided that the miner shows a leaf activity when few other understorey species

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 465 possess foliage (e.g. evergreen species, Leaves of species acclimated or adapted to high Leaf mottling wintergreen species, winter leaves of dimorphic irradiance levels generally have higher peak rates as camouflage species). Leaves of species in each these four of photosynthesis and higher levels of Rubisco and seasonal photosynthetic guilds (terminology other dark-reaction enzymes, measured per unit follows Givnish, 1987) are also exposed to rela- leaf mass, than do leaves of species grown under tively high irradiance levels and so are less likely shadier conditions (Bjbrkman, 1981); the adap- to suffer photosynthetic losses due to the reduced tive value of greater investment in leaf nitrogen leaf absorptance that accompanies mottling. content in sunnier conditions has been analysed at Furthermore, all have leaves held close to the length by Mooney & Gulmon (1979), Gulmon & ground (Givnish, 1982, 1987), in which the Chu (1981), Field & Mooney (1986) and Givnish camouflage effect should be effective. Therefore, (1986, 1988a). These models suggest a general the incidence of leaf mottling in these four guilds prediction that, among herbs of temperate deci- should be higher than in any other group of forest duous forests, the highest leaf nitrogen content herbs, or in any other growth form. should be seen among species whose leaves are It should be apparent that, in temperate deci- exposed primarily to the high irradiance, warm duous forests, evergreen herbs are especially non-limiting temperatures, and moderate to high exposed to damage by vertebrate herbivores soil water availability that prevail in spring. because they are among the few plants with Spring ephemerals fall into this category, as may a foliage available near ground level between late few early summer species whose entire leaf phe- autumn and early spring. In this regard they are nology is nearly confined to spring, or in which the joined mainly by winter annuals and wintergreen phenology of the initial leaf flush is entirely species (whose leaves emerge in autumn, overwin- confined to spring. ter and then senesce before canopy closure in the The low nitrogen/protein content of leaves is a spring) and by the winter-active leaves of dimor- fundamental constraint facing herbivores. Leaf phic species (whose leaves are borne in two nitrogen content limits herbivore digestive effici- different sets active during the cool and warm ency and growth, and strongly affects food choice seasons). Moreover, in temperate forests evergreen in natural systems (e.g. Fox & Macauley, 1977; herbs are relatively most common on infertile soils McNeill & Southwood, 1978; Milton 1979; Scriber (Givnish, 1982, 1987); replacing nutrients in & Feeny, 1979; Mattson, 1980; McClure, 1980; leaves lost to herbivores in such hibitats is espec- Lincoln et a]., 1982; Mooney & Gulmon, 1982; ially costly and should favour relatively heavy Bryant et a]., 1983; Coley, 1983, 1987; Crawley allocation to anti-herbivore defences (Janzen, 1983; Denno & McClure, 1983; Dirzo, 1984; John- 1974; Coley, 1983, 1987; Coley et al., 1985; Bazzaz son et a]., 1984; Coley et a]., 1985; Bazzaz et a]., et al., 1987). 1987). Leaf-eating herbivores are more likely to Spring ephemerals (whose leaves emerge in feed on leaves with high levels of nitrogen/protein early spring and senesce before the tree canopy than on those with lower levels (e.g. see Milton, closes) may also be particularly attractive to herbi- 1979; Coley, 1983, 1987). It thus seems reasonable vores because they have high concentrations of to assume that forest herbs with high levels of leaf nitrogen in their leaves. A study of 25 forest herb nitrogen relative to species active at the same time species in Virginia (Givnish, 1987) found that (i.e. spring ephemerals, winter-active species) spring ephemerals had the highest leaf nitrogen would be exposed to a higher potential for herbi- content (4.06 + 056%, n = 5) of any guild, vore damage than species with lower levels of leaf whereas evergreens had the lowest (2.14 ? 036%, nitrogen and thus be more likely to evolve n = 4). Spring ephemerals had a significantly camouflage. higher leaf nitrogen content than either evergreens The camouflage hypothesis advanced here bears (P < 0.001, two-tailed t-test) or early and late certain parallels to the hypothesis advanced by summer species (whose leaves emerge in spring Ehleringer et a]. (1986) to account for the extent of and reach peak cover by early to late summer visual similarity between the leaves of mistletoes before senescing) (P < 0 05, one-tailed t-test). Use and their hosts. Building on the anti-herbivore of a one-tailed test in the latter comparison is hypothesis advanced by Barlow & Wiens (1977), justified because of the expectation that spring Ehleringer et a]. (1986) argued that (poorly defen- ephemerals should have a higher leaf nitrogen ded) mistletoes with a higher level of leaf nitrogen content than summer-active species, based on than their host should visually mimic their host's their exposure to higher levels of irradiance and on leaf shape and branching pattern, so as to gain typical differences between sun and shade leaves. some measure of protection (mimetic crypsis)

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms Table 1. Growth form, leaf phenology and taxonomic affinity of vascular species with mottled leaves native to the north-eastern USA. Phenological classifications follow terminology of Givnish (1987); data drawn from Coker (1944), Fernald (1950), Mangaly (1968), Rickett (1966), Wells (1984) and Richardson, Burch & Cochrane (1987).

Growth form Leaf phenology Family Species* Notes

Forest herb Evergreen Aristolochiaceae Hexastylis virginicum, memmingeri, shuttleworthii, lewisii, arifolium, ruthii Ericaceae Chimaphila maculata Orchidaceae Goodyera oblongifolia, pubescens, repens, tesselata Ranunculaceae Hepatica acutiloba, americana Some population mottled Rubiaceae Mitchella repens Central white stripe Saxifragaceae Heuchera americana, caroliniana

Wintergreen Orchidaceae Aplectrum hyemale White veins Tipularia discolor Mottling polymorphism

Spring ephemeral Liliaceae Erythronium albidum, americanum, propullans

Early summer Berberidaceae Podophyllum peltatum Mottling fades Hydrophyllaceae Hydrophyllum appendiculatum, Spring leaves only canadense, macrophyllum, virginian um Liliaceae Medeola virginica Fruit flag Trillium cuneatum, luteum, recurvatum, sessile, viride Polygonaceae Tovara virginica Mottling fades Violaceae Viola hastata

Dimorphic Rosaceae Geum virginianum Winter leaves only Ranunculaceae Ranunculus recurvatus Winter leaves only

Woody vine Evergreen Liliaceae Smilax bona-nox Semi-evergreen Liliaceae Smilax glauca, S. lanceolata

Carnivore Evergreen Sarraceniaceae Sarracenia purpurea

Sun-adapted herb Summergreen Euphorbiaceae Euphorbia (Chamaesyce) humistrata, maculata, serpyllifolia, supina

Cases in which variegation is restricted to extremely rare genetic sports or unique, highly localized populations (e.g. Camassia scillioides forma variegata Steyermark) have been excluded. Species with leaves modified as brightly coloured floral bracts that essentially function as petals (e.g. Cornus canadensis, Monarda punctata) have also been excluded from this tabulation.

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 467 from visually oriented mammalian herbivores (e.g. lines. Graminoid genera were therefore excluded Leaf mottling phalangers). Conversely, mistletoes with lower from the survey to avoid biases caused by potential as camouflage levels of leaf nitrogen than their hosts should phylogenetic constraints or functional differences visually diverge from their host to advertise their in leaf form. Because graminoid genera are all low food value. Indeed, the degree of visual unmottled and mostly occur in open habitats, their mimicry of hosts by mistletoes does increase with exclusion is a statistically conservative treatment the relative food value of the mistletoes. Here it is that tends to bias the results against acceptance of argued that the degree of crypsis of herbs growing the first prediction of the camouflage hypothesis, in a sun-dappled understorey should be also that mottled leaves should occur more frequently related to, among other factors, the relative food in forest understoreys than elsewhere. value of different groups of herbs and the relative The ecological distribution of leaf mottling cost to them of replacing foliage consumed by among non-graminoid angiosperm genera native herbivores. to the north-eastern United States supports both predictions of the camouflage hypothesis: mottled leaves occur almost exclusively among forest herbs (Table 2) and in this group are significantly Tests of the camouflage hypothesis over-represented among evergreen and spring To test the predictions of the camouflage hypothe- ephemeral species, and among the winter leaves of sis, I surveyed the incidence of mottled leaves dimorphic species and the spring leaves of among angiosperms native to the north-eastern summer-active species (Table 3 and below). Thir- USA and adjacent parts of Canada. Botanical teen of the 20 genera with at least one species nomenclature follows Fernald (1950) except bearing mottled leaves are entirely composed of where indicated. Species with mottled leaves forest herbs and in four additional genera forest frequently occur in genera with other mottled herbs are the only mottled species present. species (Table 1), but genera within families show Mottled leaves occur in 13 of 96 genera of forest no such approach to uniformity. Consequently, herbs, compared with only seven of the remaining the incidence of mottled leaves was tallied at the 595 genera of trees, shrubs, vines, herbs of open generic level, in order to avoid overestimating the habitats and species of mixed growth forms (Table number of times the mottled leaf trait arose 2). Only two of 250 genera of herbs native to open independently while retaining maximum statisti- habitats (i.e. prairies, meadows, bogs, marshes, cal power (see Harvey & Mace, 1982; Janson, 1983). beaches, lakes, ponds, rivers and tundra) have Graminoids with narrow, linear leaves (e.g. Cyper- species with mottled leaves; only one of 36 genera aceae, Eriocaulonaceae, Gramineae, Juncaceae, of vines have mottled leaves present and none of Juncaginaceae, Xyridaceae) native to the north- the 80 genera of shrubs or 39 genera of trees have eastern USA never have mottled leaves, are excep- mottled leaves present. If open sites per se were to tionally numerous and diverge dramatically from favour mottled leaves as Smith (1986) proposed all other groups surveyed; furthermore, even if it (see his Fig. 6), one would not expect to see their were to occur, mottling of graminoid leaves would virtual absence among herbs, shrubs and trees seem to offer little potential to disrupt leaf out- with canopies exposed to sunny conditions

Table 2. Association of leaf mottling with ecological habit among non-graminoid angiosperm genera with species native to the north-eastern USA. Habitat preferences for herbaceous genera derived from descriptions by Fernald (1950): open habitats include prairies, meadows, bogs, marshes, beaches, lakes, ponds, rivers and tundra; closed habitats include forests, woodlands and thickets. Herbaceous genera with species found in both open and closed habitats were placed in the mixed category, regardless of whether single species are restricted to a single habitat class or not.

Ecological habit

Herbs of Mixed growth Trees Shrubs Vines Open habitats Forests Mixed habitats forms

Number of genera with: Mottled leaves absent 39 80 36 248 83 167 18 Mottled leaves present 0 0 1 2 13 4 0

X2= 46 9, P < 0 0001 for 6 d.f.

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 468 Table 3. Association of leaf mottling with leaf phenology among genera of forest herbs with species native to the T. J. Givnish north-eastern USA.

Leaf phenology

Spring ephemerals Evergreens/wintergreens Others

Number of genera with: Mottled leaves absent 5 5 73 Mottled leaves present 1 7 5

X2 = 24-0, P < 0-0001 for 2 d.f.

throughout much of their lives. The nearly com- sist through the summer (T. J. Givnish, unpu- plete restriction of mottled leaves to understorey blished observation). Viola hastata bears silver- herbs in the north-eastern USA flora (17 of 20 splashed leaves that senesce soon after canopy genera, 36 of 43 species) would seem to undermine closure (Klaber, 1976; S. P. Bratton, personal the general applicability of Smith's (1986) argu- communication). Podophyllum peltatum bears ment for mimicry of leaf miner damage in leaves that emerge with a mottled pattern that is Byttneria. lost later in spring; it too sheds its leaves soon after Forest herbs with mottled leaves tend to have canopy closure. Tovara virginica produces leaves leaves active in winter, or active mainly in spring marked with a dark chevron in spring, but the (Table 3). Fully 16 species are evergreen, including marking fades soon after the canopy closes even members of six genera (Chimaphila, Goodyera, though the leaves persist until fall. Geum virginia- Hepatica, Heuchera, Hexastylis [Asarum subge- num is dimorphic (Givnish, 1982, 1987), with two nus Hexastylis in Fernald, 1950] and Mitchella) in sets of leaves active at two different seasons. The six different families. Two wintergreen orchids warm-season leaves (active from mid-spring (Aplectrum, Tipularia) have mottled or striped through early autumn) are unmottled, but the leaves active from fall to spring (Table 1). Of the 12 cool-season leaves are strikingly variegated and genera composed entirely of evergreen or form an overwintering rosette. Similarly, in wintergreen herbs (Heuchera is thus excluded), Ranunculus recurvatus only the basal winter seven possess mottled leaves (Table 3). In each of leaves are mottled (H. H. Iltis, personal communi- the latter seven genera, all of the species found in cation). The sole exception to the rule associating the north-eastern USA except one (Chimaphila leaf variegation in forest herbs with photosynthe- umbellata) have mottled leaves. Leaf mottling is tic activity in winter or spring is Medeola vigi- thus especially common among evergreen or niana, which has an adaptation for fruit dispersal wintergreen forest herbs. that is highly unusual among temperate plants. Among the five understorey genera of spring Reproductive shoots bear two whorls of leaves on ephemerals that bear leaves between snowmelt an erect stem: a lower whorl of several large leaves and canopy closure, all three north-eastern species (seen on vegetative shoots too) and an upper whorl of Erythronium bear mottled leaves. Furthermore, of three smaller leaves subtending the flowers. among the remaining 15 species of forest herbs Both whorls remain entirely green from spring there is a strong association between mottling and until late summer. In early fall, the berry-like fruits photosynthetic activity before canopy closure. In ripen and turn dark blue. Simultaneously, a dia- all four species of Hydrophyllum, only the spring- mond-shaped area at the base of each leaf in the active leaves are mottled; these senesce before the upper whorl turns a striking magenta; presumably canopy closes and are replaced by unmottled this coloration is a 'fruit flag' (Stiles, 1982) and leaves that photosynthesize through the summer contrasts with the bluish berries to help attract (Morgan, 1971; T. J. Givnish, J. W. Terborgh & D. M. migrating frugivorous birds. Waller, unpublished observations). In Trillium, Thus, the predictions of the camouflage the five mottled species all belong to the same hypothesis account for most aspects of the dis- subgenus and bear sessile flowers and their leaves tribution of leaf mottling in the flora of the north- usually senesce soon after canopy closure; the eastern USA (Tables 2 and 3). Mottled leaves are unmottled species (e.g., T. cernuum, T. erectum, T. borne almost exclusively by forest herbs with grandiflorum) belong to the other subgenus, have winter- or spring-active leaves. Forest herbs that pedunculate flowers and their leaves usually per- bear individual leaves that are active only in

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 469 summer, or in both spring and summer, generally E. citrinum, E. helenae, E. hendersonii, E. Leaf mottling lack variegation: the only exceptions are Tovara howellii, E. multiscapoideum, E. oregonum, E. as camouflage virginiana, in which the mottling pattern of indi- revolutum) tend instead to grow in lower, more vidual leaves does fade after canopy closure; and closed sites, such as forests, woodlands and Medeola virginiana, in which variegation seems chaparral (Rickett, 1970; Munz, 1975). clearly related to fruit dispersal. A few short evergreen, or semi-evergreen, species of forest vines in the genus Smilax also have mottled Alternative hypotheses leaves, as might be expected in species that are ecologically rather analogous to evergreen forest Several alternative explanations advanced by herbs. Leaf mottling does not occur in herbs of Smith (1986) for the incidence of leaf mottling open sites, except in the special case of the have already been critiqued (see above). In review- carnivorous Sarracenia purpurea, and in the ing this report, A. P. Smith (personal communica- puzzling case of four mat-forming species of tion) suggested that leaf mottling in evergreen and Euphorbia. In some species of Euphorbia, red wintergreen forest herbs, spring ephemerals and coloration occurs in leafy bracts that help attract the spring leaves of early summer species may pollinators; the appearance of minute red dots or serve to reduce leaf temperature and transpiration, midribs in certain local species (each with minute in accord with his hypothesis (iii), based on the leaves) may represent a non-adaptive vestige of a fact that these plants are exposed to relatively high floral adaptation in an ancestral population. (It is irradiances and low soil temperatures during fall, interesting to note that in the exceptional genera in winter and/or spring. Although it is possible that which animal attraction is apparently involved leaf mottling has such an additional beneficial [Medeola, Sarracenia, Euphorbia], the leaves are effect, it seems unlikely that such an effect would mottled red, not white, pale green or black as in all be substantial given the relatively high humidity other cases). Leaf mottling is completely absent in near the forest floor. Furthermore, if leaf energy shrubs and trees of the north-eastern flora. Even balance were a significant factor contributing to though woody plants are exposed to high ir- the adaptive value of mottled leaves, the question radiance levels they should lack mottled leaves, would still remain as to why plants with such because they (i) do not experience dappled light leaves should not instead have uniformly more environments in forest understoreys throughout reflective foliage. their lives; (ii) are so tall that they could not If leaf energy balance were generally important 'disappear' against a sun-dappled litter layer even in favouring leaf mottling in forest herbs, then if their leaves were mottled; and/or (iii) do not have mottling should be common and of high reflec- the unusual leaf phenologies favouring leaf mott- tance (i.e. whitish) in plants with broad leaves, ling that are found in many understorey herbs. because they are the most liable to high leaf The distribution of leaf mottling in the genus overtemperatures and/or transpiration rates Erythronium provides good support for the pre- (Gates, 1962, 1980; Parkhurst & Loucks, 1972; dictions of the camouflage hypothesis and the idea Taylor, 1975; Givnish & Vermeij, 1976; Gates, that sun-dappled forest understoreys, not open 1980; Geller & Smith, 1980). In addition, mottling sites per se, favour leaf mottling. Sixteen species should be more common in herbs of deciduous occur in North America and each has a short forests, because the heat loads they face in winter period of photosynthetic activity early in the and early spring would be greater than those in the growing season (Fernald, 1950; Rickett, 1970; understorey of evergreen forests. Munz, 1975). As noted earlier, three species are The first of these two predictions is not sup- spring ephemerals with mottled leaves, found in ported by the data presented in this report. Among deciduous forests in eastern North America (Table forest herbs native to the north-eastern USA and 1). The remaining species range from the Rocky adjacent parts of Canada, there are a total of 13 Mountains to the Pacific coast and include species genera with mottled leaves (Table 2). Of these, with mottled or unmottled leaves. Those with dark mottling (brown, black or purple blotches) unmottled leaves (E. grandiflorum, E. klama- occurs in four of the seven genera with the thense, E. montanum, E. purpurascens, E. tuolum- broadest mottled leaves (Erythronium, Podo- nense) occur primarily in open high-elevation phyllum, Tipularia, Tovara). Lighter, more reflec- sites, such as alpine and subalpine meadows, or tive leaf patterns appear in broad-leaved species of openings in subalpine forest (Rickett, 1970; Munz, Aplectrum, Hexastylis and Trillium, and in all but 1975). Those with mottled leaves (F. californicum, one of the remaining six genera characterized by

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 470 Table 4. Classification of genera of forest herbs, based on presence (0) or absence of leaf mottling and on whether they T. J. Givnish generally occur in forests with (+) or without (-) evergreen trees frequent in the canopy. Genera for which the available information is inadequate (or whose species are too variable in distribution) to make the latter classification are indicated by the absence of a trailing + or -, and have been excluded from further statistical analyses (see text).

Liliaceae Ranunculaceae Limnanthaceae Hydrophyllaceae Arisaema - Aconitum - Floerkea - *Hydrophyllum - Clintonia + Actaea - Elatinaceae Nemophila - Convallaria + Anemonella - Bergia - Boraginaceae Disporum - Cimicifuga - Violaceae Cynoglossum - *Erythronium - Coptis + Hybanthus - Mertensia - Maianthemum + *Hepatica + Onagraceae Labiatae *Medeola - Hydrastis - Circaea Blephilia - Polygonatum - Isopyrum - Araliaceae Collinsonia - Smilacina - Berberidaceae Panax Cunila - Streptopus - Caulophyllum - Umbelliferae Meehania - * Trillium - Diphylleia + Cryptotaenia - Micheliella Uvularia - Jeffersonia - Erigenia - Scrophulariaceae Orchidaceae *Podophyllum - Osmorhiza - Scrophularia - *Aplectrum - Papaveraceae Sanicula - Seymeria - Calypso + Dicentra - Pyrolaceae Orobanchaceae Corallorhiza Sanguinaria - Moneses + Epifagus - * Goodyera + Cruciferae Monotropa Conopholis - Hexalectris Dentaria - Monotropsis + Acanthaceae Isotria - Saxifragaceae Pterospora + Dicliptera Ponthieva - Astilbe - Pyrola Rubiaceae * Tipularia - Boykinia - Ericaceae *Mitchella + Triphora - Mitella *Chimaphila + Phrymaceae Urticaceae Tiarella - Epigaea + Phryma Laportea - Rosaceae Gaultheria + Caprifoliaceae Aristolochiaceae Aruncus - Primulaceae Linnaea + Asarum - Dalibarda + Trientalis + Triosteum Aristolochia - Waldsteinia Loganiaceae Compositae *Hexastylis - Leguminosae Spigelia - Adenocaulon + Polygonaceae Amphicarpa - Gentianaceae Chrysogonum * Tovara - Desmodium Halenia + Elephantopsis Portulacaceae Phaseolus - Asclepiadaceae Polymnia Claytonia - Gonolobus Polemoniaceae Polemonium

smaller effective leaf sizes among species or sea- excluded from the analysis; all such genera had sonal morphs with mottled leaves. Because dark unmottled foliage. No significant difference in the mottling is likely to increase, not decrease, leaf incidence of mottling among herbaceous genera temperature and/or transpiration, it should be less native to deciduous vs evergreeen forests was frequent than light mottling in broad-leaved detected (X2 = 0.45, P < 0 5 for 1 d.f.). The same is species if leaf energy balance were generally true if only mottled taxa with light splotches on important in helping to favour leaf mottling. The their foliage are included in the analysis (X2 = distribution of dark vs light patterns of mottling as 2 13, P < 0 7 for a one tailed-test with 1 d.f.). Given a function of effective leaf size does not support the lack of empirical support for both corollaries of this hypothesis. Smith's energy balance hypothesis, it seems The second corollary of Smith's leaf energy reasonable to conclude (at least preliminarily) that balance argument is also not supported by the it is not of general applicability among temperate distribution of forest herbs with mottled leaves. forest herbs. Genera composed solely of forest herbs were classified into those that generally occur in habi- Relation of leaf mottling to other proposed tats with few or no evergreen trees in the canopy vs mechanisms of crypsis in flowering plants those that often or usually inhabit forest with evergreen trees present (Table 4). The few genera The role of leaf mottling as camouflage in forest for which inadequate data were available were herbs joins a growing list of proposed mechanisms

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 471 of vegetative crypsis in angiosperms. Such background based on their small difference in Leaf mottling mechanisms include (i) visual mimicry of stones parallax and/or focal plane relative to the ground. as camouflage and pebbles by stone-plants (Lithops and several Indeed, even in By. aculeata vertebrate herbi- other members of the Aizoaceae, Asclepiadaceae, vores may play a role in fostering the evolution of Crassulaceae, Euphorbiaceae and Liliaceae) in mottled leaves. A. P. Smith (1986; personal com- South African deserts, and by certain Cactaceae munication) observes that leaf mottling is present (Ariocarpus, Pediocactus and Sclerocactus) in only in vines relatively close to the ground (< 1 m North American deserts (Purpus, 1914; Marloth, tall) and disappears in older, taller plants. Inter- 1929; Woodruff & Benson, 1976; Barlow & Wiens, estingly, short variegated plants tend to have more 1977; Wiens 1978); (ii) resemblance of dead grass thorns/stem length than taller or unvariegated by spines of certain cacti that typically grow plants (Smith, 1986). Thorns are usually inter- among clumps of particular grass species (Wiens, preted as a mechanical defence against vertebrate 1978); (iii) mimicry of dead sticks by leafless herbivores and tapirs have been observed to con- succulents in arid and semi-arid regions of sume foliage of B. aculeata (Smith, 1986). The southern and eastern Africa (Wiens, 1978); (iv) presence of a mechanical anti-vertebrate defence visual mimicry by mistletoes of the leaf size, shape in plants with a potential visual defence and the and texture of better-defended and/or less decrease or disappearance of both defences in nutritive foliage of their hosts (Barlow & Wiens, plants not accessible to (or, in the case of the visual 1977; Ehleringer et al., 1986); (v) convergence on defence, not effective on) vertebrate herbivores, common leaf shapes by plants defended with an provides indirect support for the camouflage uncommon (and hence, dissimilar) chemical hypothesis. A. P. Smith (personal communication) defence (Rausher, 1981); (vi) visual divergence offers the additional corroboratory observation among species with similar chemical defence, to that leaf mottling in a common tropical under- thwart visually orienting herbivores capable of storey herb in Panama, Calathea inocephala, dis- overcoming that defence (Gilbert, 1975); (vii) appears in plants taller than 80 cm. visual mimicry of various vegetative stages of Why are mottled leaves relatively common in human crop plants by weeds (Barrett, 1983); (viii) tropical forest understoreys? Although it is im- the active folding movement of the tender, bright possible at present to draw any definite con- green foliage of the creeping Mimosa pudica in clusions, one possiblitiy is that mottled leaves are response to mechanical stimuli (perhaps via con- favoured because many tropical soils are tact with a grazer's mouth or feet), leading to their extremely infertile, favouring short evergreen almost complete visual disappearance against a herbs (Givnish, 1982) and making any loss of background of pasture grasses and herbs (T. J. foliage to herbivores relatively costly, particularly Givnish, unpublished observation); (ix) apparent given the low levels of light prevalent in many mimicry of dead leaves by the dull brownish tropical forests. Such unproductive conditions foliage of certain, short-statured palms of tropical should favour relatively heavy allocation to anti- Malaysian forest understoreys (Stone, 1979); and herbivore defences (Coley, 1983, 1987; Coley et al., (x) the brownish or blackish leaves seen in many 1985) and the low stature of many tropical under- tropical understorey herbs (Blanc, 1989). storey herbs should render leaf mottling a viable Blanc (1989) proposes that the latter act as mechanism of crypsis. Corollaries of this hypothe- camouflage against the dark exposed soil of tropi- sis would be that (i) leaf mottling should be cal forest floors and notes that almost all such common in herbs of relatively short stature; and leaves are held within 20 cm of the ground, where (ii) leaf mottling should be relatively common in they would be most visually effective against tropical montane forests or on extremely infertile vertebrate herbivores. Blanc (1989) also observes parent materials and less common on rich volcanic that almost all of the tropical forest herbs with soils (except under extremely shady conditions). mottled leaves he surveyed hold their leaves within 20 cm of the ground, but draws no inference Conclusions regarding the adaptive significance of such leaf mottling. The fact that mottled leaves in both The camouflage hypothesis presented in this temperate and tropical forests are usually held report appears to account for most aspects of the close to the ground lends additional support to the distribution of leaf mottling in the flora of the camouflage hypothesis, because such leaves north-eastern USA. As predicted, mottled leaves would be both accessible to vertebrate herbivores are borne almost exclusively by short forest herbs and capable of visually 'disappearing' into the with winter- or spring-active leaves, are especially

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 472 common in evergreen species and are virtually camouflage mechanism proposed here seems to T. J. Givnish absent in herbs of open sites, vines, shrubs and yield considerable insight into the basis for leaf trees. The hypothesis can potentially be extended mottling, coupling adaptation to irradiance level to account for the relative abundance of mottled with a context-specific anti-herbivore defence. As leaves among tropical forest herbs, and parallels in Neger (1913) noted, the puzzling and oft-over- certain respects other recently proposed mechan- looked phenomenon of leaf mottling provides a isms of crypsis in angiosperms based on leaf size, rich field for further physiological and ecological shape, texture, colour and/or movement. studies. It should be clear, however, that additional independent tests of the camouflage hypothesis Acknowledgments are needed to corroborate the findings presented here. Particularly useful would be data on rates of I wish to thank S. P. Bratton, H. H. Iltis, S. P. vertebrate and invertebrate attack on mottled vs Solheim and T. S. Cochrane for providing helpful unmottled leaves, nutrient levels and chemical observations and references, and S. E. Knight, K. L defences in mottled and unmottled leaves, photo- Steudel, C. E. Umbanhowar and D. M. Waller for synthetic light response in both types of leaves, useful discussions. A. P. Smith provided several and the incidence of leaf miner damage as a constructive comments and suggestions in his function of growth form and leaf phenology. review of the original manuscript. Observations and field experiments involving herbivory on mottled and unmottled seasonal leaf morphs of Hydrophyllum might be especially References rewarding in this regard. Barlow, B. A. & Wiens, D. (1977) Host-parasite resem- It should be emphasized that leaf mottling may blance in Australian mistletoes: the case for cryptic arise through more than one mechanism: the leaf mimicry. Evolution, 31, 69-84. miner damage mimicry hypothesis of Smith (1986) Barrett, S. C. H. (1983) Crop mimicry in weeds. Economic seems quite plausible for the case of Byttneria and Botany, 37, 255-282. his suggestion that mottling may sometimes serve Bazzaz, F. A., Chiariello, N. R., Coley, P. D. & Pitelka, L. F. (1987) Allocating resources to reproduction and as aposomatic coloration should not be over- defense. BioScience, 37, 58-67. looked, particularly for species with a bold, visu- Benzing, D. H. (1980) Biology of the Bromeliads. Mad ally striking (i.e. non-dappled) pattern of River Press, Eureka, California. variegation. In some plants, vivid leaf markings Bj6rkman, 0. (1981) Responses to different quantum may serve as floral attractants, as in several tropi- flux densities. In Physiological Plant Ecology. L Responses to the Physical Environment (eds. O. L. cal members of the Gesneriaceae (Morley, 1973; Lange, P. S. Nobel, C. B. Osmond & H. Ziegler), pp. Stiles, 1982). This last possibility could account 57-107. Springer-Verlag, New York. for the sole instance of leaf variegation in the flora Blanc, P. (1989) Biologie des plantes des sous-bois of south-western Australia, Hakea victoriae. The tropicaux. These de Doctorat d'Etat, Universit6 Pierre et Marie Curie (Paris 6), Paris. leaves of this shrub, patterned brilliantly in orange Bryant, J. P., Chapin III, F. S. & Klein, D. R. (1983) and yellow, surround inconspicuous flowers Carbon/nutrient balance of boreal plants in relation to (Erickson et al., 1979) and may serve to attract vertebrate herbivory. Oikos, 40, 358-368. honeyeaters and other avian pollinators that visit Coker, W. C. (1944) The woody Smilaxes of the United them, perhaps in concert with the plant's un- States. Journal of the Elisha Mitchell Society, 60, 27-69. branched and unusually tall growth form in the Coley, P. D. (1983) Herbivory and defensive character- low heaths in which it occurs (cf. Givnish et al., istics of tree species in a lowland tropical forest. 1986). In a few carnivorous species, reddish and Ecological Monographs, 53, 209-233. yellow leaf mottling may serve to mimic flowers Coley, P. D. (1987) Interspecific variation in plant anti- and attract potential prey (Givnish, 1988b). herbivore properties: the role of habitat quality and Benzing (1980) suggests that the dark leaf bases of rate of disturbance. New Phytologist, 106 (suppl.), 251-263. certain epiphytic tank bromeliads may help pro- Coley, P. D., Bryant, J. P. & Chapin III, F. S. (1985) vide a microsite in which various animals might Resource availability and plant anti-herbivore conceal themselves from predators and perhaps defense. Science, 230, 895-899. result in crucial inputs of nutrients via animal Crawley, M. J. (1983) Herbivory: The Dynamics of Animal-Plant Interactions. Blackwell Scientific Pub- excretions. Finally, there remain instances of leaf lications, Oxford. mottling (e.g. the spotted orchis Dactylorhiza Denno, R. F. & McClure, M. S. (eds.) (1983) Variable fuchsia of European meadows and grasslands) Plants and Herbivores in Natural and Managed Sys- with no obvious explanation. Nevertheless, the tems. Academic Press, New York.

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 473 Dirzo, R. (1984) Herbivory: a phytocentric overview. In Current Problems in Sociobiology (ed. King's College Leaf mottling Perspectives on Plant Population Ecology (eds. R. Sociobiology Group), pp. 343-361. Cambridge Univer- sity Press, New York. as camouflage Dirzo & J. Sarukhan), pp. 141-165. Sinauer Press, Sunderland, Massachusetts. Janson, C. H. (1983) Adaptation of fruit morphology to Ehleringer, J. R., Ullmann, I., Lange, O., Farquhar, G. D., dispersal agents in a neotropical forest. Science, 219, Cowan, I. R., Schulze, E.-D. & Ziegler, H. (1986) 187-188. Mistletoes: a hypothesis concerning morphological Janzen, D. H. (1974) Tropical blackwater rivers, animals, and chemical avoidance of herbivory. Oecologia, 70, and mast-fruiting in the Dipterocarpaceae. Biotropica, 234-237. 6, 69-103. Erickson, R., George, A. S., Marchant, N. G. & Morcombe, Johnson, N. D., Chu, C. C., Ehrlich, P. R. & Mooney, H. A. M. K. (1979) Flowers and Plants of Western Australia. (1984) The seasonal dynamics of leaf resin, nitrogen A. H. & A. W. Reed, Sydney. and herbivore damage in Eriodictyon californicum Evanari, M. (1989) The history of research on white- and their parallels in Diplacus aurantiacus. Oecolo- green variegated leaves. Botanical Review, 55, 106- gia, 61, 398-402. 139. Klaber, D. (1976) Violets. A. S. Barnes & Co. Fernald, M. L. (1950) Gray's Manual of Botany. Van Lincoln, D. E., Newton, T. S., Ehrlich, P. R. & Williams, K. Nostrand, New York. S. (1982) Coevolution of the checkerspot butterfly Field, C. & Mooney, H. A. (1986). The photosynthesis- Euphydryas chalcedona and its larval food plant nitrogen relationship in wild plants. In On the Diplacus aurantiacus: larval response to protein and Economy of Plant Form and Function (ed. T. J. leaf resin. Oecologia, 52, 216-223. Givnish), pp. 25-55. Cambridge University Press, New McClure, M. S. (1980) Foliar nitrogen: a basis for host York. suitability for elongate hemlock scales, Fiornia Fox, L. R. & Macauley, B. J. (1977) Insect grazing on externa (Homoptera: Diaspididae). Ecology, 61, 72-79. Eucalyptus in response to variation in leaf tannins and McNeill, S., & Southwood, T. R. E. (1978) The role of nitrogen. Oecologia, 29, 145-162. nitrogen in the development of insect/plant relation- Gates, D. M. (1962) Energy Exchange in the Biosphere. ships. In Biochemical Aspects of Plant and Animal Harper & Row, New York. Coevolution (ed. J. B. Harbourne), pp. 77-98. Gates, D. M. (1980) Biophysical Ecology. Springer-Ver- Academic Press, London lag, New York. Mangaly, J. K. (1968) A cytotaxonomic study of the Geller, G. N. & Smith, W. K. (1980) Leaf and environ- herbaceous species of Smilax: section Copros- mental parameters influencing transpiration: theory manthus. Rhodora, 70, 55-82; 247-273. and field measurements. Oecologia, 46, 308-313. Marloth, R. (1929) Stone-shaped plants. South African Gilbert, L. E. (1975) Ecological consequences of a co- Biological Society, 6, 1-8. evolved mutualism between butterflies and plants. In Mattson, W. J. (1980) Herbivory in relation to plant Coevolution of Animals and Plants (eds. L. E. Gilbert & nitrogen content. Annual Review of Ecology and P. H. Raven), pp. 210-240. University of Texas Press, Systematics, 11, 119-161. Austin. Milton, K. (1979) Factors influencing leaf choice by Givnish, T. J. (1982) On the adaptive significance of leaf howler monkeys; a test of some hypotheses of food height in forest herbs. American Naturalist, 120, selection by generalist herbivores. American Natura- 353-381. list, 114, 362-378. Givnish, T. J. (1986) Economics of biotic interactions. In Mooney, H. A. & Gulmon, S. L. (1979) Environmental and On the Economy of Plant Form and Function (ed. T. J. evolutionary constraints on the photosynthetic char- Givnish), pp. 667-680. Cambridge University Press, acteristics of higher plants. In Topics in Plant Popu- New York. lation Biology (eds. 0. T. Solbrig, S. Jain, G. G. Johnson Givnish, T. J. (1987) Comparative studies of leaf form: & P. H. Raven), pp. 316-337. Columbia University assessing the relative roles of selective pressures and Press, New York. phylogenetic constraints. New Phytologist, 106 Mooney, H. A. & Gulmon, S. L. (1982) Constraints on leaf (suppl.), 131-160. structure and function in reference to herbivory. Givnish, T. J. (1988a) Adaptation to sun and shade: a BioScience, 32, 198-206. whole-plant perspective. Australian Journal of Plant Morgan, M. D. (1971) Life history and energy relations of Physiology, 15, 63-92. Hydrophyllum appendiculatum. Ecological Mono- Givnish, T. J. (1988b) Ecology and evolution of carnivo- graphs, 41, 329-349. rous plants. In Plant-Animal Interactions (ed. W. A. Morley, B. C. (1973) Ecological factors of importance to Abrahamson), pp. 243-290. McGraw-Hill, New York. Columnea taxonomy. In Taxonomy and Ecology (ed. Givnish, T. J. & Vermeij, G. J. (1976) Sizes and shapes of V. H. Heywood), pp. 265-281. Academic Press, New liane leaves. American Naturalist, 100, 743-778. York. Givnish, T. J., McDiarmid, R. W. & Buck, W. R. (1986) Fire Munz, P. A. (1975) A California Flora and Supplement. adaptation in Neblinaria celiae (Theaceae), a high-ele- University of California Press, Berkeley. vation rosette shrub endemic to a wet equatorial tepui. Neger, F. W. (1913) Biologie der Pflanzen. Enke, Oecologia, 70, 481-485. Stuttgart. Gulman, S. L. & Chu, C. C. (1981) The effect of light and Parkhurst, D. F. & Loucks, 0. L. (1972) Optimal leaf size nitrogen on photosynthesis, leaf characteristics, and in relation to environment. Journal of Ecology, 60, dry matter allocation in the chaparral shrub, Diplacus 505-537. aurantiacus. Oecologia, 49, 207-212. Prokopy, R. J. & Owens, E. D. (1983) Visual detection of Harvey, P. H. & Mace, G. (1982) Comparisons between plants by herbivorous insects. Annual Review of taxa and adaptive trends: problems of methodology. In Entomology, 28, 337-364.

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms 474 Purpus, J. (1914) Mimikry bei Kakteen. Mueller's Deuts- Smith, A. P. (1986) Ecology of a leaf color polymorphism T. J. Givnish che Garten-Zeitung, 29, 73-78. in a tropical forest species: habitat segregation and Rausher, M. D. (1981) The effect of native vegetation on herbivory. Oecologia, 69, 283-287. the susceptibility of Aristolochia reticulata (Aristolo- Stiles, E. W. (1982) Fruits flags: two hypotheses. Ameri- chiaceae) to herbivore attack. Ecology, 62, 1187-1195. can Naturalist, 120, 500-509. Richards, P. W. (1952) The Tropical Rain Forest: an Stone, B. C. (1979) Protective coloration of young leaves Ecological Study. Cambridge University Press, New in certain Malaysian palms. Biotropica, 11, 126. York. Taylor, S. E. (1975) Optimal leaf form. In Perspectives in Richardson, J. W., Burch, D. & Cochrane, T. S. (1987) The Biophysical Ecology (eds. D. M. Gates & R. B. flora of Wisconsin preliminary report No. 69. Euphor- Schmerl), pp. 73-86. Springer Verlag, New York. biaceae - the spurge family. Wisconsin Academy of Wells, E. F. (1984) A revision of the genus Heuchera Sciences, Arts and Letters, 75, 97-129. (Saxifragaceae) in eastern North America. Systematic Rickett, H. W. (1966) Wild Flowers of the United States, Botany Monographs, 3,45-121. vol. 1. McGraw Hill, New York. Wiens, D. (1978) Mimicry in plants. Evolutionary Bio- Rickett, H. W. (1970) Wild Flowers of the United States, logy, 11, 365-403. vol. 6. McGraw Hill, New York. Woodruff, D. & Benson, L. (1976) Changes of status in Schultz, J. C. & Baldwin, I. T. (1982) Oak leaf quality Sclerocactus. Cactus and Succulent Plants Journal, declines in response to defoliation by gypsy moth 48,131-134. larvae. Science, 217, 149-151. Scriber, J. M. & Feeny, P. (1979) Growth of herbivorous caterpillars in relation to feeding specialization and to Received 7 March 1989; revised 25 January 1990; growth form of their food plants. Ecology, 60,829-850. accepted 7 February 1990

This content downloaded from 130.223.51.59 on Wed, 27 Apr 2016 13:23:21 UTC All use subject to http://about.jstor.org/terms