Search

Studies in hazel undergrowths in five Minnesota pine stands provided techniques for estimating total hazel biomass and its three components (leaves, aerial stems, underground parts). Regression equations were developed for estimating total biomass and the three components based on measurements of aerial stem weight, aerial stem basal area, or number of aerial stems. Nutrient concentrations in the three biomass components were not well enough correlated with biomass nor with each to allow estimation using regression techniques. Total hazel biomass ranged from 4,600 to 15,000 kg/ha representing 4% to 7% of the total above ground weight of the stands. Compared to the overstory, hazel appeared to be rich in Ca, P. and Mn and contained from 2% to 22% of the weight of the nutrients in the above—ground parts of the vegetation. See full-text article at JSTOR

Mark-recapture studies were made on a grid over an extended period of time to investigate home range utilization by small mammals, especially Peromyscus leucopus, the most numerous small mammal on the study area. Since captures were largely concentrated around a home range center, Calhoun's (1963) bivariate normal density, which predicts a more diffuse capture pattern, did not adequately describe space utilization in Peromyscus. Mark-recapture results showed a similar pattern as a continual removal regimen: 50% of the total Peromyscus captured in the first 30 days were taken before the first Blarina and Pitymys. This seems contrary to Calhoun's (1963) predictions that as "dominant" species are removed (here Pero- myscus), "subordinate" species (here Blarina and Pitymys) enlarge their home ranges and are caught more often. My results can better be explained by invoking behavior-mediated differ- ences in trappability between species and/or ecological differences in the use of the habitat (fossorial vs. runways in undergrowth, etc.). Finally, my evidence suggests that Peromyscus are arranged in "family" groups consisting of one adult female, several adult males (one dominant, several subordinate), and a number of juveniles.

Values of forest transpiration rate, calculated from a vapor diffusion model that uses the vapor pressure defict (v.p.d.) of the canopy air, and measurements of the stomatal resistance (every 2—3 h) and leaf area index of the canopy, agreed well with those obtained from energy balance/Bowen ratio measurements. Stomatal resistance characteristics were also used in the diffusion model instead of actual stomatal resistance measurements to calculate transpiration rate. There was reasonable agreement between these transpiration values and energy balance measurements averaged over 3—h periods. Both the model, using stomatal resistance characteristics, and energy balance measurements showed that transpiration rate increased with increasing v.p.d. until a certain v.p.d. was reached after which the rate declined. The values of v.p.d. at which maximum transpiration rate occurred were 1.5, 1.3 and 1.1 kilopascals [=15, 13 and 11 mb] for soil water potential ranges of zero to —350 kilopascals [=0 to —3.5 bars], —350 to —950 kPa [=3.5 to —9.5 bars] and —950 to —1250 kPa [=9.5 to 12.5 bars] respectively. Calculations from the model, using stomatal resistance measurements (every 2—3 h) on 7 days, indicating that the fraction of transpiration from the thinned stand transpired by the salal undergrowth during daytime ours increased from approximately 45—70% during a 4—week drying period.

Both physical and biological processes shape species assemblages (communities). For birds, vegetation structure has long been assumed to be the dominant factor in habitat selection, especially along successional gradients. While vegetation may be important as a proximate factor, detailed knowledge of ultimate factors governing habitat selection is required. Gradients of microcli- mate, especially temperature and moisture, may be such an ultimate factor through direct physiolog- ical pressures on birds or indirectly through distribution and availability of food resources. We doc- umented the existence of gradients of both vegetation structure and microclimate in the undergrowth of seasonally humid forest in central Panama. To assess the relative importance of these gradients in shaping local avian distribution, birds were netted in undergrowth (up to 3 m above ground) during 2-wk periods in dry (March) and wet (July) season for 4 yr (1979-1982). A total of 3037 captures of 95 species was recorded during that period. Although patterns of species richness and capture rate are relatively simple, species composition and other assemblage attributes are complex and difficult to interpret without careful evaluation of the dynamics of individual species. The variety of activity patterns exhibited by birds on moisture and vegetation gradients clearly demonstrates the complexity of dynamics that affect assemblage attributes. Overall, birds are more active at dry sites and at sites with intermediate shrub density. However, activity levels change with time, suggesting that avian activity reflects a dynamic process of habitat selection. Though no food resource data are available at present, the general pattern of habitat use on diurnal time scales for several guilds suggests that birds track microclimatic optima for physiological reasons. Habitat selection processes are both ex- tremely complex and variable in time and space for the avifauna of tropical forest undergrowth. Indeed, the scale of study in space and time is important in determining the conclusion of a study. The recent perception of a dichotomy between equilibrium (deterministic) and nonequilibrium (stochastic) assemblages is undermined by these results. Habitat selection in birds produces dynamic (nonequilibrium) assemblages in space and time. However, these are not stochastic assemblages. Each species seeks habitat optima in the context of current environmental conditions on diurnal, seasonal, and between-year time scales. Thus, instead of nonequilibrium and stochastic assemblages, the avifaunas of tropical forest undergrowth are nonequilibrium but relatively predictable from knowl- edge of current environmental conditions.

The role of hazel shrub and herbaceous undergrowth layers in nutrient cycling was studied in two red pine (Pinus r esinosa Ait.) and two paper birch (Betula papyrifera Marsh.) stands on similar soil types in northern Minnesota. The annual litterfall produced by these undergrowths and by the overstory was sampled at 15 points in each stand as was the forest floor; the top 3 cm of mineral soil was sampled at 7 of the points. In the pine stands both undergrowth layers significantly added to the total dry weight of litterfall. Under hazel there was significantly more N, Ca, and Mn, and under herbs there was more than twice the amount of K in the litterfall than where these undergrowth layers were absent. In the birch stands the differences in the nutrient content of the litterfall beneath these undergrowth layers were not great, and undergrowth layers had less effect on the nutrient cycle. Although greater amounts of nutrients occurred in the litterfall under hazel and herbaceous layers, the amounts of total nutrients in the forest floor under each undergrowth type were not significantly different nor were total N, available P, or exchangeable K, Ca, Mg, and Mn in the top 3 cm of mineral soil. Turnover rates were different under each type of undergrowth and in red pine stands the forest floor under hazel and herbs appeared to decompose more rapidly than the forest floor of pure pine, but not as rapidly as the forest floor under either hazel or herbs in the birch stands.

Four islands, 8 to 13 miles f rom the mainland, separate the Mississippi coast from the Gulf of Mexico. These include Cat Island on the west, and Ship Island, Horn Island, and Petit Bois on the east. Horn Island is in Jackson County; the west end of the island, the longest one in the group, is 13 miles southeast of Biloxi, Mississippi, at approximately 300 13' N. and 88? 40' WV. The climate of Horn Island is mild; records available at Pascagoula, Mississippi, some 12 miles northeast, show a mean winter temperature of 52.9? F. and a mean summer temperature of 80.70 F. The highest temperature recorded at Pascagoula is 1010 F. and the lowest 160 F. The mean annual precipitation is about 58 inches, most of which falls during the warmer part of the year. Geologists believe that these coastal islands were at one time a part of the mainland, and that the entire Gulf coast is slowly sinking (McGee, 1891). The islands are separated from the mainland by a depression, which has become filled by the waters of the Gulf of Mexico, forming the Mississippi Sound. Soil samples taken at different places on Horn Island seem to substantiate this hypothesis, for in a number of places the soil is quite similar to that of the mainland. The nature of the bottom of the Sound, which is composed of hard clay, and the shallowness of the Sound, further substantiate this hypothesis of the origin of the islands and the sinking of the coast. Horn Island is 13 miles long and threefourths of a mile wide at its widest part. At the west end of the island is a sand bar approximately 2 miles long and at the east end another bar about one-half mile long (fig. 1). Between these sand bars the island is heavily wooded with slash pine.' Along the south (or Gulf) and north (or Sound) coasts, dune and beach sands form the major soils. In the interior of the island, are many depressions, some of which are inhabited by dune vegetation and others by marsh plants. In both conditions, Plummer fine sand predominates. This soil consists of fine, dark gray sand underlain by fine, white, water-soaked sand. This poorly drained soil supports scattered grass, which affords grazing for a hundred or more cattle. The sand bars at each end of the island are practically barren, the vegetation consisting mainly of widely scattered tufts of grass, with small patches of Croton and Hydrocotyle. Towards the interior of the island the ground rises to about 10 feet above sea level, and slash pine, together with an undergrowth of shrubs constitutes the major vegetation. On a few scattered patches on old dunes prickly pear occurs. At both the west and the east ends of the island sand dunes much higher than in the interior have buried tall pine trees to their crowns. Slash pine, the principal forest species, forms pure stands of various ages; the oldest trees measure 26 inches in diameter at breast height (d.b.h.) and are about 170 years old. Trees 12 inches or more d.b.h. are 90 to 100 years old, and those 8 to 10 inches d.b.h. probably are 40 to 50 years old. The reproduction ranges

Extensive forests of quaking aspen (Populus tremuloides Michx.), which occur widely at intermediate elevations in the Intermountain and Rocky Mountain West, constitute an important resource, both esthetically and economically. The aspen type is esteemed among sightseers, picnickers, and campers because of its beauty. Aspen forests protect the soil of an important snow-collecting belt on western mountain watersheds. Although aspen is not one of the important timber species, aspen wood products are in considerable and increasing demand. Finally, the type supports undergrowth capable of furnishing a great amount of cover and forage for wildlife and livestock. On aspen range in good condition one may wade waist deep through a rich mixture of many plants, including species of Heracleum, Mertensia, Delphinium, Osmorphiza, Agastache, Erigeron, Rudbeckia, Senecio, Thalictrum, Agropyron, Bromus, Elymus, and Carex. Most aspen range in the Intermountain region is depleted from prolonged overgrazing, however, and palatable species that were formerly abundant are likely to be scarce or absent. Curiously, much less vegetation is produced in openings, as a rule, than beneath the aspen canopy.3 Usually the vegetation of openings is the sparser and shorter, and it tends to include a smaller proportion of desirable forage species (Houston 1954). It may, indeed, include undesirable species such as the annual Madia glomerata Hook., that are absent under the aspens. We have personally observed such differences in production and species composition on aspen ranges in Utah, Nevada, southern Idaho, western Wyoming, and western Colorado. These observations have included a great variety of sites with respect to soil and exposure, and many variations in character of vegetation. Why should such a difference in ground-cover production exist? Is it a product of those factors. responsible initially for the existence of openings in an aspen forest? Is it caused by differences in microclimate under and away from the aspen canopy? Is it a result of heavier grazing in openings than under the aspen? Or is it the result of interactions between some or all of these factors? Whatever the cause, this widely observed difference seems anomalous, because it would appearthat demands upon the environment by the aspen trees themselves might be expected to infringe on the needs of plants of the shrub and herb layers. Logically, therefore, being free of such competi-tion, the openings should produce more vegetation, not less, than equal areas under aspens. Such a relation has been described by Moinat (1956) in the Quercus gambelii type in southwestern Colorado, a zone somewhat warmer and drier than the aspen-fir zone. Moinat shows that production ofthe field layer in grassy parks is markedly higher than in scrub-oak thickets. Many trenching experiments (reviewed byKorstian and Coile 1938) have demonstrated the adverse effects of root competition by forest overstory on growth of native ground vegetation or artificially planted trees. Shirley (1945), in a study carefully designed to test the relativeeffects of shade and root competition of aspen anct jack pine on tree seedlings and planting stock in Minnesota, demonstrated a root-competitive depression of conifer growth, although not of sur-vival. Shirley's comment is that, "In these studies, however intense the root competition of the overwood, its effect on survival was more than offset by the benefits of the shade, provided this did not reduce the light intensity below 20 percent." Weknow of no experiments of this kind involvingwestern aspen except for Pearson's (1914) comparisons of survival of planted Douglas-fir under aspen and in openings. These showed better sur-vival under aspen.

The mutual hostility of ants belonging to colonies of different or even of the same species is so nearly universal, that it is surprising to find colonies belonging to different taxonomic subfamilies actually living in the same nest in a state of mutual toleration or even amity. Prof. Auguste Forel was the first to describe such an occurrence and to designate it as parabiosis, to distinguish it from those peculiar forms of social symbiosis, called "compound nests " and " mixed colonies," among certain Nearctic and Palearctic species. While on a trip to Colombia in i896 he made the following observations (i898): "I often observed, for the first time in the neighborhood of Santa Martha, two species of ants belonging to different genera and even to different subfamilies, a Dolichoderus and a Crenastogaster, both of a shining black color, the former much the larger and especially broader than the latter, and of a very different form, running very generally in common files, both over the ground and on the undergrowth. The files were very long and dense, so that the ants met and elbowed one another continually. " The two species were foraging on the bushes, the Cremastogaster seeking especially the plant-lice or scale insects, the Dolichoderus especially the juices of plants. Hence towards their terminations the files divided, each species proceeding to its own feeding ground. I finally succeeded in discovering on the trunk of a mango a large termite nest which had been appropriated by the two species of ants in question and was serving them as a habitation of a character hitherto unknown. The time for observation was unusually propitious, as each species had its sexual phases and its pupae in the nest. The latter was inhabited just as it had been abandoned by the termites, without any alterations or additions. In no part of it were the two species of ants actually mixed. Some of the corners were still occupied by the termites. " The chambers and galleries were nearly everywhere occupied either by the Cremastogaster with their females, males and pupae or by the Dolichoderus with their females, males and pupae. Each species, therefore, had its own separate menage, unlike the conditions seen in the mixed formicaries of our PolYergus and Formica, which carry on their households in common.

Invasion of red pine stands in northern Minnesota by beaked hazel occurred mainly by seeding, and dense undergrowths were formed by a coalescence of many individual clones (about 2,000/acre). In young undergrowths, where beaked hazel was becoming established, 25—43% of the individuals were in the seedling stage under 12 years old, and 72—92% of the clones were under 20 years old. Clonal development, which is slow and limited, begins with the initiation of underground stems when seedlings are 7—12 years old. From 46% to 73% of the clones were only 30 cm in diameter or under, and the largest clone (over 30 years old) occupied an area of only 1.2 by 1.8 m. The control of beaked hazel undergrowth made be possible by light burning while the plant is in the seedling and young clonal stage and by controlling seed source. See full-text article at JSTOR

Considerable interest has been directed to the study of the chemical composition of tree leaves. It was discovered long ago that species differ in their abilities to extract nutrients from the same kind of soil, and that the minerals are not similarity disposed in the tissues, with a result that their litter varies considerably in composition. The silviculturist sometimes finds it desirable to maintain in a mixed stand certain trees that produce low-value wood, only because they are useful in maintaining soil fertility for associated valuable species. The speed of rehabilitation of waste-lands and eroded areas varies with the species used in reforestation. Game management must often take into account the nutrient content of foliage available for browsing. Some analyses have been made in the hope that a reliable method of evaluating soil fertility could be discovered. Although nutrient content often appears unrelated to soil fertility as indicated by chemical analyses of soil or by vigor of growth, special correlations of local significance have found with sufficient frequency that this relationship always deserves study. In any inquiry into why trees grow where they do, the varied influences of the different species must be given careful consideration, for small initial differences in environment may cause differences in the pioneer vegetation, then the contrasted floras add so much more difference as to obscure the primary factor that caused initial divergence (Chandler 1939). Despite the importance attached to the subject, only indirect information was available for just one species of the tree flora of the northern Rocky Mountains, when the present study was undertaken. In summarizing North American literature on the chemical composition of litter of about 60 trees that have been analyzed, Kittredge (1948) listed PopuIts tremuloides, studied in other parts of the continent, as being low in N (in the group with 0.4-0.8%)) medium in P (0.150.3% ), medium in K (0.3-0.6%), and medium in Ca (1.5-3.0%c). Tarrant, et al. 1951 have recently published analyses of fresh leaves and of weathered litter for ten trees along the Pacific coast, four or five of which are taxonomically identical with trees included in this report. Since the chemical composition of foliage varies progressively throughout the summer, and their report does not indicate the time of year when the foliage collections were made, their data cannot be compared with the present study. Although the primary objective of this study has been to evaluate species influences upon soil through the character of the litter they produce, several subsidiary problems were recognized. With the fundamental forest associations of eastern Washington and northern Idaho defined (Daubenmire 1952), material could be collected according to habitat types and thus make a preliminary exploration of the possibility that chemical composition may be related to variations in forest undergrowth (the Finnish "forest site types") and rate of tree growth (the North American "site index"). Also, the phytosociologic study showed two major discontinuities in forest undergrowth. A rich xerophytic undergrowth in the lower forest zones

The main type of forest humus layer known as mull (Muller) seems, as a rule, conditioned by a fauna efficient in keeping down the fungi and maintaining a more bacterial type of decomposition (Muller, 1887; Falck, '23: and Romell, '34). In addition, this fauna may effect an extensive mixing of humus with mineral soil. This is most pronounced in extreme forms of mull such as the crumb mull (Romell and Heiberg, '31 ), inhabited by large earthworms. At least outside Ramann's country, the r6le of earthworms in creating and maintaining a crumb mull type of humus layer has been rather generally accepted after Muller's classical studies (1887). Recently, this has been realized by Ramann's own pupil and successor, Lang ('31). On the other hand, the agents responsible for the formation of less extreme mull forms have remained obscure. Observations reported in the present paper suggest a r6le in the formation of certain mull forms to large myriapods of the American diplopod genus Fonttaria Gray. The findings are believed to have some interest also for animal ecology. Prof. J. W. Bailey, of the University of Richmond, Va., has kindly informed me that most, if not all, work done on myriapods has been systematic in nature, with just a mention, here and there, of feeding habits, etc. In localities with detritus mull (Romell and Heiberg, '31), large diplopods have repeatedly attracted the attention of Professor Heiberg and the writer and particularly of Mr. C. Heimburger, during his extensive field studies in the Adirondacks (Heimburger, '33). Specimens collected by Hfeimburger were kindly determined by Charles H. Blake, of the Boston Museum of Natural History, and by Charles E. Johnson, of the Roosevelt Wild Life Station at the N. Y. State College of Forestry, Syracuse, N. Y. They were found to belong to two species, FoWtaria triniaculata (Wood) and Fonltaria coriacea Koch (see Bollman, p. 123). Both species were described from northeastern United States by Wood, the latter one under the species name corrugatufs. In one particular locality, near North Hudson in the eastern Adirondacks, the largest species, F. triniacuilata, wvas found to be very plentiful and exerting a most striking soil effect. The site is a steep south slope carrying a stand of middle-aged white pine with hardwood undergrowth and a ground vegetation of Heimburger's Dicentra type on a sandy loam rich in lime and showing a brown soil profile. More detailed data are given by Heimburger ('33) under

In a previous paper (4, I9I7) the writer has described forest succession in Puget Sound sphagnum bogs, and in other papers (4, I9I3; I9I4; I9I9) has listed the trees growing in various sphagnum bogs of the Puget Sound region and Alaska. All of these are comparatively early stages in bog succession. None of them have reached a stage where the forest is dominant. An area of about three acres near Victoria, British Columbia, shows a very late stage where the forest has become the dominant vegetation, and sphagnum is no longer found living. This area was visited in July, 1920, and a more careful study was made during several visits to it at various times in July, I92I. The area is at present bordered on the south by a meadow, on the east and north by Lost Lake, and on the west by the Patricia Bay line of the Canadian National Railway. The forest consists of a pure stand of lodge-pole pine, Pinus contorta. The trees are of good size for this species, many of them being 20 to 30 feet tall and 6 to io inches thick at the base. A few are 12 inches thick. The stand of trees is very dense over nearly all of the area, and many young trees have been killed by the shade of the taller ones. The trunks of these trees are straight and slender. They do not have the scrubby form that this species does on rocky, exposed places, nor the stunted appearance that it has in the early stages of forest succession in Puget Sound bogs. Very few pine seedlings were found, and these only on the border along the west side, where the forest was cleared a few years ago when the railroad was built, and Nature was then allowed to take her course again. Lodge-pole pine is common on this portion of Vancouver Island. It is found around lakes, both in swampy places and on rocky shores, and on the thin soil of rocky hills. It is found practically to the top of Mount Benson, near Nanaimo. Glendenning (i, i9i8) says that it is frequent in bogs and on gravel ridges in the Cowichan district on Vancouver Island. Whitford and Craig (5, i9i8) state that this species is general throughout the province of British Columbia, from the international boundary to the Yukon drainage, being found at an altitude of 5,ooo to 6,ooo feet, and attaining a height of 50 to 75 feet and a diameter of I2 to 24 inches. The undergrowth in the bog forest under consideration consists of a dense thicket of Labrador tea, Ledum qroenlandicum, and salal, Gaultheria shallon.

Samples based on stand measurements and borings of trees, and clippings of undergrowth, were taken from 24 forests and forest heaths in the Great Smoky Mountains. Samples were taken also from heath balds and grassy balds and a California coastal redwood forest. Available information on ratios and regressions of forest tree and shrub dimensions were used to estimate stand biomass and net annual production above ground. Mature climax forests of mesic environments below 1,400 m are characterized by: wood basal areas of 50—64 m2/ha and basal area increments of 0.3—0.6 m2/ha/yr, stem wood volumes (parabolic estimate) of 750—900 m3/ha and estimated wood volume increments of 530—590 cm3/m2/yr, aboveground biomasses of 500—610 t/ha and aboveground net annual productions of 1,000—1,200 g/m2, and biomass accumulation ratios of 40—50. These and other stand dimensions decrease along the moisture gradient to xeric sites and decrease toward higher elevations. Above—ground net annual productions of forest heaths of xeric slopes and forests of highest elevations are 420—650 g/m2. Evergreen spruce—fir forests are more productive than deciduous forests above 1,400 m. Among mesic high—elevation beech and fir forests, production is higher on south slopes than on north slopes. Production and biomass of steady—state forests were estimated from multiple correlations using elevation and weighted—average index of site moisture as independent variables. Apart from some high—elevation stands, production is not significantly correlated with evergreen vs. deciduous foliage or with direction of exposure affecting incident sunlight. When unstable stands are excluded, production and biomass are strongly correlated with each other and with mean tree height and other stand dimensions. A wide range of temperate—zone climax forests of relatively favorable environments have net annual productions above and below ground of 1,200—1,500 g/m2. Productions of stable closed heath balds are lower, 700—1,200 g/m2, but productions of heath balds and less productive forests overlap broadly. Production of unstable forests, both successional stands and those opened by removal of large trees, exceeds that of steady—state forests of similar environments. Canopy coverage and light penetration are not strongly correlated with forest production. Light penetration to the herb stratum ranges from 1.4—7.0% of incident sunlight in more open forests and forest heaths to 0.3—0.9% in cove forests. Foliage live/dry ratios decrease along the moisture gradient from mesic to xeric stands–from about 5.0 to 2.8 in shrub clippings, 7.6 to 2.8 in herb clippings. Undergrowth production and biomass are trivial compared with the tree stratum in many forests. Shrub production is generally higher in xeric environments and is 20—145% of tree production in forest heaths. Herb production is higher at the extremes of the moisture gradient (exceeding 3% of total aboveground production in mesic and in open xeric stands) than in intermediate stands (below 1%). Apart from such differences, her and thallophyte biomass and production increase with elevation to maximum values in fir forests of the highest summits.

A community in which the southern white cedar, Clhaiaecyparis tlyo ides (L.) BSP., is dominant occurs on peat soils of the Atlantic coastal plain. Such f orests are neither common nor extensive in view of the enormous areas of peat land in this region. The most complete study of southern white cedar is that by Korstian and Brush ('31). This study, together with those of previous investigators, supplied ecological information such as site characteristics, the aspect of the community, associated tree species, and some features regarding reproduction. The site occupied by the southern white cedar community is consistently peat bog underlain by sand (Korstian, '24; Korstian and Brush, '31). Korstian has pointed out that, with larger proportions of silt and clay in the underlying mineral soil, swamp hardwoods are more abundant. Hardwood-cypress swamps are underlain by clay or sandy-clay subsoil. The aspect of the southern white cedar forest has been described and illustrated adequately. Gifford ('00) gave a description of the community in which he emphasized the extreme density of the forest, the trees growing so close together that they afford mutual support and the bases of the crowns meeting to form a solid canopy. He mentioned the dense undergrowth, but neither he nor any of the other workers has analyzed its composition except to note that it consisted of a dense growth of vines and shrubs. Although Harper ('26) listed associated species of all categories, trees, shrubs, lianas, herbs, and bryophytes in order of their abundance in his account of a middle Florida cedar swamp, the cedars were not the largest nor most conspicuous trees in this community. It must therefore be considered as a mixed bog forest which cannot be compared with the true cedar community. Although Korstian and Brush ('31 ) found that pure, dense, more or less even-aged stands are most characteristic f or the southern white cedar, they listed numerous arborescent associates. The number of associated tree species is greater in the southern part of the range than in the northern part. It is apparent from the literature that southern white cedar reproduction requires decidedly open conditions. Pinchot ('00) noted that no reproduction occurs after crowns meet to form a continuous canopy. Korstian and Brush ('31) likewise found that seedlings do not tolerate the shade of the cedar forest. They pointed out that, although the seedlings may start under cover of bog shrubs, the shrubs usually develop rapidly into an impenetrable jungle making establishment of the cedar difficult. Ackerman ('23) found that seedlings come up in abundance in the rollways and tramways after logging, but that elsewhere the slash and the rank growth of shrubs and vines produce a mass which the young cedars have small chance of penetrating. This situation was recognized likewise by Korstian and Brush ('31). It is significant in this connection that in New Jersey the southern white cedar constantly invades cranberry bogs (Gifford, '00). Recognizing the requirement for an open seed bed, at least one free from the extreme competition of the bog shrubs and lianas, several authors suggested clearing to obtain reproduction. In view of the dangers involved in controlled burning in such areas, Ackerman ('23) suggested the use of cat

Theories of changing climate are fascinating to the student of history; seemingly conclusive on the surface, they are apt to prove inconclusive when subjected to a strict examination of facts. Cape Henlopen's famous wave of sand seemed clearly to point to a physiographic change of wind components; but an examination showed that it was due to destruction of timber and undergrowth which had held great dunes in situ. Similarly anyone who has seen the area covered by Golden Gate Park, San Francisco, with an interval of half a century between visits, would be justified in attributing the transformation to a change in the distribution of rainfall. But there was change neither of wind nor of rainfall. In every physiographic region there are species which live on the ragged edge of a precarious existence. A slight change in mean temperature or in the distribution of rainfall is the signal for their disappearance. The change may be temporary but the effects may be permanent and profound. Even within our own land the removal of protective shrubbery has been followed by destructive corrosion, leaving a given area in a condition which made reclamation impossible. A change in the efficiency of a people; a change in the demand for a commercial product of the soil; the invasion of a highly productive region by warlike intruders-all these are historical examples that have reduced fertile regions to conditions approaching those of a desert. The most productive areas in the world have been made so by irrigation. Irrigated lands are deficient in rainfall. The soil of such lands is usually very rich. It is rich because excessive rains have not leached the nutrition therefrom, thus reducing its producing power. When the cultivation of such lands ceases, they quickly revert to conditions apparently irreclamable. The foregoing does not apply to the changes in temperature which are indicated by geological data, such as the glacial periods. The evidence concerning these is fairly conclusive; in one instance it may be regarded as positive. The various glacial epochs, however, have no relation to the climatic changes which are alleged to have occurred in historic and near-historic times. Within the periods of written history there is no trustworthy evidence of progressive changes either in temperature or in rainfall occurring to productive lands. There are erratic changes, it is true, and it is probable that Bruckner has established evidence of an oscillatory period. These changes

Chestnut blight, caused by the Ascomycete Endothia parasitic, was first discovered in the United States in 1904 in New York City and spread rapidly northeast, west, and southwest. By 1910 it had spread as far south as southeastern Pennsylvania and in 1920 the "80 per cent infection zone" extended along the Blue Ridge mountains to central Virginia. In 1926 local infections had been reported in western North Carolina, northern Georgia, and northwestern South Carolina. In 1930 Gravatt and Gill reported that infection in western North Carolina showed eighty to ninety-nine per cent. Owners of chestnut timber in the newly infected areas of the Carolinas and Georgia quickly cut as much of that lumber as possible. Most of the remaining chestnut (Castanea dentata)lt trees, dead or alive, have been gradually removed, until now, twenty years after the blight reached its peak, there are only a few scattered ghost trees left to indicate the once relatively great importance of the species. The removal of a species which made up more than 40 per cent of the overstory trees in the climax forests of an area would certainly be expected to produce changes in the composition of the vegetation of the region. It is possible that species co-dominant with chestnut would advance in rank and become the most important species in the new climax. It is also possible, as Braun (1950) suggests, that the opening of the canopy, the release of undergrowth trees, and the resulting changes in the humus layer may cause

There is no doubt that forest undergrowth can be put to practical use in forestry, not only for the simple purpose of stand description but also for the interpretation of successional trends, regeneration possibilities, site quality, and so forth. The chief obstacle to the fuller use of the minor species is lack of knowledge about them-their distributions, habits, tolerances, and hence the significance of their presence or absence at any particular place. For this reason it has been common practice when attempting to make some use of the forest flora to ignore the strictly ecological approach. In the study of forest site types, for example, the usual method adopted has been to search for recurring patterns of plant species in the hope that these can be be related to the occurrence and growth of particular forest types, and although satisfactory results have apparently been obtained in some instances, confusing or negative results have been obtained in others, with the result that plant indicators have acquired a bad name. It would be foolish to condemn the concept because it turns out to be less simple than at first suspected; the obvious need is for a more critical approach, and a closer study of the forest plants. It is safe to say that as more is learned of the ecology of the individual species, the more they will be found useful for various purposes. The present article outlines a method by which some of the more obvious "indications" of plants may be utilized in forestry. The basic factors on which life depends are light, warmth, moisture, and nutrients, and it is the variability of these over the surface of the earth that is primarily responsible for the variability of plant and animal communities. The study of the ecology of a natural community consists largerly of an analysis of the roles played in community organization by the basic factors, with particular attention to those which are believed to be limiting, as for example, temperatures in the arctic and moisture in desert regions. Unfortunately, it is not always possible to appraise directly the quantity and quality of the basic necessities of life in a given situation, and recourse to indicators is the only solution. The forester, interested in an assessment of moisture, nutrients

The Third International Phytogeographic Excursion, which had been proposed in i9i5 by the Swiss botanists, and which had been postponed on account of the World War, was held during the summer of 1923. The main party met in the Botanical Museum of the Technical High School in Zurich and was addressed by the head of the High School, the rector of the University, and the committee of arrangements composed of Drs. C. Schroeter, H. Brockmann-Jerosch and Ed. Ruibel. After the lectures, delivered in French and German, the party inspected the finely appointed laboratory of botany with its important collection of teaching material. The first excursion was taken on July 21st to study the Sihlwald, with midday meal at the Hotel Hirschen Albis, while the afternoon was spent in a visit to Utliberg and Uto-Kulm; July 22nd was spent near Zurich in a visit to the mixed forest of Loorenkopf and Adlisberg, with an inspection of the forest experiment station at Adlisberg under the guidance of the director of the station. A two days' excursion with return to Zurich on the evening of July 24 was taken via Brunnen, Luzern and Alpnachstad to the summit of Mt. Pilatus, where the night was spent, the return down the mountain being made on foot the next day. Zurich, which had been the headquarters during the first five days of the excursion, was finally left on July 25. The party entrained for Chur and in rapid succession until August 13, when the party disbanded at Kleine Scheidegg at the foot of the Jungfrau, the points of principal ecologic and phytogeographic interest were visited in Switzerland and Northern Italy. Only some of the important features of the excursion can be mentioned within the limits of this synopsis. At Chur-Reichenau xerotherm vegetation was studied. The party proceeded by Albula Pass into the Engadin, traveling from Bevers to Zernez in the Lower Engadin. Zernez is the gateway to the Swiss National Park, which was visited, as well as the summit of the Buffalora, which commands a view of the Ortler in Tyrol. The botanists found the flora of Alp Buffalora and Mt. Buffalora, just outside the National Park limits, rich and interesting. The forest vegetation in the Upper Engadin between St. Moritz and Pontresina was found to consist of a mixture of Larix decidua, Picea excelsa, Pinus silvestris var. engadensis and woody undergrowth of Vaccinum11 Myrtillus, V. uliginosum, Rhododendrumsn ferrugineuuw, B etula pubescens, Pyrus aucuparia, etc.

In considering the vegetation which becomes finally established in any given region, one finds that the present method is to designate such growth as the climax vegetation. The use of the term climax in this connection seems to be misleading, since the word climax conveys the impression of an abundant or copious vegetation, whereas the present usage of the term refers to a final or ultimate condition, regardless of mere vegetational display. The dense shade of the ultimate forest results in an undergrowth consisting of strictly shadetolerant plants; such a forest floor frequently presents a paucity of vegetation. It is obvious that a vegetation of this character may be less luxuriant than that found in previous stages where light permits a dense undergrowth; the climax in vegetational display may be thus attained previous to the ultimate condition. It seems that the word " ultimate" possesses many advantages over the term " climax " when used to designate the final vegetation of a region. Objection might be raised to the use of ultimate on the grounds that no one knows what the ultimate condition will be eventually; however, in an ecological sense the usage of the word is directed toward present-day ultimate conditions in much the same manner as the term climax is now generally applied.2 The need of a term to designate any member of the ultimate vegetation of any given region seems apparent; at the present time such a plant is usually referred to as a " climax-plant," a " member of the climax vegetation " or by other more or less similar forms of expression. As a term of convenience, the word eschatophyte is proposed as one applicable to any member of an ultimate vegetation. For example, in various parts of the Appalachian Mountains the Maple-Hemlock formation is always found in the ultimate vegetation; such plants may be called eschatophytes. On a large area of Canada, various species of conifers are eschatophytes. It should

Newly created oceanic islands slowly accumulate species. In contrast, habitat islands created by rising water level or by human reduction of the area of a habitat (for example, by deforestation) support a full complement of local species at their formation. With time these islands lose species, a phenomenon called relaxation (Diamond 1972, Wilcox 1980). Several explanations may account for relaxation in island relicts. (1) As island size decreases, diversity of habitats present decreases. Species dependent entirely or in part on lost habitats disappear ("habitat diversity hypothesis;" Able and Connor 1979, Wilcox 1980). (2) Islands passively intercept dispersers in proportion to their area. As island size declines, recolonization probabilities decline. Thus, species lost from a small land bridge island are less likely to recolonize ("disperser intercept hypothesis;" Able and Connor 1979, Martin 1980, Wilcox 1980). (3) Smaller islands have more species with low population sizes, a condition which predisposes such species for extinction from random causes ("rarity hypothesis"; Willis 1974, Terborgh and Winter 1980). Corollaries of the "rarity hypothesis" relate bird size and ecological specialization to extinction probability (Willis 1974). (4) On theoretical grounds, species unable to maintain stable populations in the face of environmental vagaries (species with high "population variability") may be prone to local extinction (Leigh 1975, 1981). Terborgh and Winter (1980) suggested from qualitative considerations that species dependent on resources likely to be variable and/or patchy in their distribution (especially frugivores and nectarivores) are prone to extinction. Tracking of those resources in space results in locally variable population sizes for such species. To my knowledge, no quantitative field data are available to test the "population variability hypothesis." To what extent do these hypotheses account for loss of species on Barro Colorado Island (BCI), Panama, a 14.8-km2 land bridge island formed when the Panama Canal was created by damming the Chagras River early in the 20th century? I have used two approaches to explore that question. First, I have examined in detail the avifauna of BCI and its presumed source fauna in the adjacent mainland forest of the Pipeline Road region of Parque Nacional Soberania. Soberania is a 22 000-ha tract of lowland forest on the mainland adjacent to BCI. At its nearest point, BCI is ;200 m from a peninsular extension of that forest, although less-disturbed forest is not so close. As many as 5060 species of forest birds are missing from BCI (Karr 1982), well above earlier conclusions that 15-18 forest species are extinct on BCI (Terborgh 1974, Willis 1974, Wilson and Willis 1975, Willis and Eisenmann 1979). The approach used in earlier estimates of avian extinctions on BCI led to underestimation of extinction rates. Further, the high extinction rate seems to be due to the restricted habitat mosaic of BCI relative to that found on a similar-sized mainland area. Undergrowth and ground species and species associated with foothill forest are especially prone to extinction (Karr 1982). Second, I have conducted a long-term study of undergrowth species in mainland forest near BCI, with the objective of identifying population attributes of species prone to extinction. I report those findings here. Birds of forest undergrowth have been sampled periodically since 1968 with mist nets in an area of 2 km2 along the Pipeline Road (centered at Limbo Hunt Club [LHC]) and 9 km east of BCI. A total of 4722 captures of 114 species of birds is available in samples collected from August 1968 through March 1982. The methodology for this study was operation of 12-14 mist nets in the undergrowth of forest for periods of 3-6 d (see Karr [1979, 1981] for details of methodology). All data were collected by me except for 10 monthly samples (826 captures) from 1977 to 1978 collected by Schemske and Brokaw (1981). Net shyness sometimes occurs in areas that are regularly sampled with mist nets. However, sampling in recent years has been limited to two 3-6 d periods each year, and.there is no sign of net shyness with this low sampling frequency. Indeed, 5070W of captures are of birds banded during earlier sampling periods. In this paper I concentrate on two of the four hypotheses described above: rarity and population variability. I do not consider the disperser intercept hypothesis since it is incompatible with the empirical observation that the undergrowth forest birds discussed in detail here are very reluctant to cross water gaps. The habitat diversity hypothesis will be discussed only briefly here as it relates to population variability since it is the primary focus of another paper (Karr 1982). I assume that the relative abundances of species today in the vicinity of Limbo Hunt Club are similar to those of the bird community of BCI at the time BCI became an island.