A quantitative description of some coast forests of Natal

A lthough the s tands are com plex and are seemingly heterogeneous, there are definite patterns of species behaviour, and trends a ie indicated for the 10 stands and for the four forest areas. Relatively few species are d o m in an t in each s tand an d the ap p a re n t diversity is mainly due to the high percentage of species that are relatively u n c o m m o n in each s tand. The m eth o d s described in this paper should be applicable in a study of a b roader range o f N a ta l forests, an a rea from which quanti ta t ive studies have heretofore been virtually absent. I N T R O D U C T I O N Eew quanti ta t ive studies o f subtropical forest vegetation have been m ade in South Africa. These forests have been described subjectively and profiles have been draw n to show the different layers. More specifically, virtually no quanti ta t ive studies have been made of the coasts forests in Natal until recent studies in which Moll (1968a; 1968b; 1969) gathered density and frequency da ta for various synusiae in large quadrats . On the other hand, rapid techniques to sample large area o f forest vegetation have been used in the temperate zone, for example the quar te r method (Cottam and Curtis, 1956), where quanti ta t ive analytic techniques have been developed (Curtis, 1959). In temperate zone forests, however, the vegetation is more hom ogenous and the num ber o f w oody species is smaller com pared to subtropical forests. The quarter method was briefly investigated in Natal forests, but was not considered suitable because it requires a larger hom ogenous area than was found to exist. Instead o f the quar te r method, this paper describes an area-m ethod o f sampling, which was applied in 10 stands o f forest in N ata l , and the subsequent quanti ta t ive analysis o f these stands.

Eew quantitative studies of subtropical forest vegetation have been made in South Africa.These forests have been described subjectively and profiles have been drawn to show the different layers.More specifically, virtually no quantitative studies have been made of the coasts forests in Natal until recent studies in which Moll (1968a;1968b;1969) gathered density and frequency data for various synusiae in large quadrats.
On the other hand, rapid techniques to sample large area of forest vegetation have been used in the temperate zone, for example the quarter method (Cottam and Curtis, 1956), where quantitative analytic techniques have been developed (Curtis, 1959).In temperate zone forests, however, the vegetation is more homogenous and the number of woody species is smaller compared to subtropical forests.The quarter method was briefly investigated in Natal forests, but was not considered suitable because it requires a larger homogenous area than was found to exist.Instead of the quarter method, this paper describes an area-method of sampling, which was applied in 10 stands of forest in Natal, and the subsequent quantitative analysis of these stands.

S T U D Y AREA
The 10 stands of forest, from four areas of Coast Forest, were studied in December 1967.Table I gives a summary of various physical characteristics of the stands.Stands I to 3 were located in the Krantzk I oof Nature Reserve, 25 km northwest of Durban, and are shown as area A on the map of Natal (Fig. 1)-Stand I was on a Table Mountain Sandstone slope above a cliff.Stand 2 was just below the 100 m high cliff, and stand 3 was closer to the valley bottom than stand 2 (Fig. 2).The underlying rocks of stands 2 and 3 were granite, but many large sandstone rocks  I acli staiul was situated well within the forest to a\oid any edge eflcct.I lie total forest m each of the lour areas ranges from 50 to 250 ha.
I he lorests disetisscd here are classified by Acocks (1953) as ( oasial I ropical forest, although the krant/klool stands are at the inland and altitudinal limits of this tvpe.I dwards (1967) classifies the forests m this part of Natal from sea level to 500 m as C oast 1 ow lands forest, and from 500 to I 200 m above sea level as Semi-Coast forest.
I he four forest areas discussed in this paper are representative of the few remaining remnants of a once much more extensive and perhaps continuous forest along the coast dunes, in the lower valleys, and on east-and south-facing slopes inland to about the 5(H) m contour level.Approximately 90 percent of this forest type has been destroyed since the arrival of Bantu and I uropean settlers in the last 400 to 500 vears (Brookes and Webb. 1965).
I he best climatic data available for the stands are statistics from Durban (Anonymous, 1954).The mean dailv temperatures are 24 C in February and 16.5 C in July, with an annual mean daily temperature of 20.5 C. The absolute minimum and maximum temperatures are 4 C and 42 C. respectively.Light frost is known to occur occasionally in the Krantzkloof Forest, but the other forests may be considered frost-free.Average precipitation for Durban ranges from 2S mm in Juls to 121 mm in November, with a vearlv average ot I 00 N mm.These precipitation fig u re s do not take into account the high humidity throughout the sear.resulting in heavy dews in winter.

C ( )M IM I \ I
The number of species of vascular plants in a region gives an indication of the potential complexitv of its plant communities.South Africa has .inestimated 16 (M X) to IS 000 species of which some 5 (XX) species occur in Natal (Ross, 1972).Six hundred to 650 of the Natal species are woody, i.e.. trees, shrubs and lianes (Ross. 1972: Ross andMoll. 1972).
Natal lias rugged and diverse topography, rising from the Indian Ocean to over 3 300 m on the Drakensberg Range within a distance of 160 km.The main slopes and aspects, and the diverse climatic conditions associated with mountainous terrain lead to m am microclimates.With the many species from which to select, the plant communities are extremely complex Since Natal forests are virtually unstudied quantitavelv.and vel occur in such an interesting situation regarding topography, climate, and numbers of woody spccies.it was considered desirable to attempt a quantitative studs.Flic 10 stands of Subtropical Coast Forest studied probably represent some of the most complex forests in Natal.If rapid quantitative methods could be devised which work in this region, a broader comparative study of the forest vegetation of Natal could be undertaken.The aim of these quanti tative methods was to sample and describe a larger .ire.i of forest, rather than to describe only one plot at a time, as has usually been done in studies of tropical and subtropical forest.A further aim was to assess the degree of complexity and heterogeneity.The samples consequently represented the vegetation of one or two hectare, which is still much smaller than those Curtis (1959) and his associates sampled in the temperate forests of Wisconsin where their samples usually represented vegetation of 5 to S hectare.
1 ll l l) Mi I MODS Flic following criteria were followed in the selection of stands.T he stands were in upland situations free from standing water at any time of the year, and were free from any recent disturbance by man or domestic animals.
I lie prime criterion for selection was topographic homogeneity: an area of one to two hectare m si/e with the same slope, aspect and soil parent material was selected.The vegetation had "visual homogeneity" from a structural standpoint, in that the age or si/e classes of species, the physiognomy, and the density were relatively uniform throughout the stand.It was also determined that certain of the prominent tree species did occur throughout the stand.However, an actual estimate of vegetational homogeneity could not be determined until analysis of the samples.There was no attempt to study the full range of successional stages and all the stands sampled were relatively mature.
Each stand was sampled with five 1/10 acre circular plots.A I 10 acre plot has an area of 4 356 square feet, and therefore a radius of 37,2 feet."The worker who recorded the data (Rogers) stood in the centre of the plot and determined the outer perimeter of the circle using a field rangefinder.A rope marked with the exact radius was also used in dense situations, and for trees located near the perimeter.The other worker (Moll) identified the species, and measured them with the aid of an assistant, Bernard Mkhize.
Two categories of woody species were recorded.Saplings and large shrubs were those individuals having a diameter at breast height (4.5 feet) of one to four inches.Trees, measured with a basal area tape giving readings in square inches, were those individuals having a d.b.h. of more than four inches, i.e., a basal area of more than 12 square inches.The basal areas of occasional trees witi buttresses at 4.5 feet were measured above the butresses.Basal area of the tree trunk is a common measure of size or dominance (Curtis, 1959), and is an easier and more objective measure than height or canopy spread in a closed forest.Herbs, lianes.epiphytes and small shrubs were not sampled.
The information recorded in the field for each stand, therefore, was altitude, slope, aspect, soil parent material, and measurement or listing of woods species of over one inch in d.b.h.In addition, a soil sample was collected in each of the five plots per stand: these were then combined to give one soil sample for each stand.)RESULTS A N D DISC USSION It svould have been possible to assess each tree regarding its presence in the canopy or other sub ordinate layers.However, whether a tree is in the canopy or not is relative and varies with site, topo graphy, age of the stand, which other species are present and interpretation of the observer.Therefore, canopy individuals were characterized arbitrarily as those having a basal area of 100 square inches or more.This si/e was not only convenient, but gave a reasonable separation.It was empirically determined in the stands that trees exceeding this si/e class formed the canopy layer a remarkable number of times.Trees having a basal area of less than 100 square inches were classified as subcanopv.Those of less than 12 square inches were listed as understorey, a category which includes saplings of canopy and subcanopy trees and large shrubs.Using these criteria we found a good correlation in our study with the same species subjectively assigned to the various layers by Henkel (1934) andMoll (1967).
In Table 2. the data recorded for the canopy, subcanopy and understorey species are shown for various layers.The order in which the stands are Iisteel is based on an ordination which will be discussed later.Canopy, subcanopy and understorey layers are designated by C, S and U respectively.Various totals and averages for layers as well as percentage relative dominance are also shown.(oaSdi cirCti).X signifies that a species was present in the stancL_but"was not recorded,^ the ^m ple.Not included in this table are 24 species which occurred in one standjonly, although thev are included in the totals.These species are listed in the appendix.For each of the three categories of species the dominant in each laver in each stand, is indicated by parentheses ■ J '   The number of trees per aere in the canopy ranged from 26 to 82 in the 10 stands.Trees per acre for the canopy species in the subcanopy layer ranged from 22 to 134, while trees per acre of subcanopy species in the subcanopy layer ranged from eight to 180, etc.

Species
Summary totals and averages for the layers, and for each stand as a whole, arc listed at the bottom of Table 2.The total number of species appearing in the various layers ranges from 20 to 40 species per stand.

Basal area as a measure o f relative dominance
Relative dominance, based on basal area figures, is used as a measure of importance because it bears a fairly direct relationship to the amount of light intercepted or shade cast, the amount of water and nutrients taken up, photosysthesis, litter dropped, etc., which are underlying reasons for attempting to assess importance of the individual species.
Note then at the bottom of Table 2 that although the total number of stems per acre ranges from 268 to 1 116 (with an average of 658 • 60%), the total basal area per acre ranges only from 16 438 to 24 514 square inches (an average basal area of 19 714 • 20°,,).This same pattern was found in a study of 86 stands of deciduous forest in southern Wisconsin which showed all degrees of cutting from heavy to light (Rogers, 1959), where the number of stems per acre ranged from 231 to 1 454, while the average basal area was 18 391 4 25%.It appears that there may be some phenomenon similar to "carrying capacity" operating in temperate and subtropical hardwood forests, i.e.. as a forest matures from n large number of small stems to a smaller number of larger stems, the amount of ground actually occupied by the trees does not change nearly as much as the number of trees.Furthermore, the approximate average of 20 (XX) square inches is measured at breast height.Measurement of the basal area of a tree at ground level is usually a little over twice as great as at breast height, i.e., a basal area figure at ground level would approximate 40 000 square inches per acre.With approximately 6,3 million square inches to an acre, this means that less than one per cent of the ground is actually occupied by the trunks of saplings and trees.These statistics are mentioned to illustrate that there may be some biological principle involved in basal area figures.If this is true, the use of basal area figures as an indication of relative dominance for species would be more than a convenience, and, in fact, might have some absolute significance.
The canopy species vary in relative importance from 61% in stand 9 to 97% in stand 7.In five of the 10 stands (1,9,10,7,8).canopy species dominate in the subcanopy layer.Only in one stand (stand 1) do understorey species, i.e., shrubs, dominate the understorey layer.In every other stand it is the subcanopy species which predominate in the under storey layer.Using basal area as an indication of importance, the canopy layer is shown to be the dominant layer in all but stands 1 and 9.If more stands covering a broader successional range of forests were studied, it is possible that these relative dominance values could be used to indicate successional trends.On the basis of relative dominance of canopy layers and canopy species, for example, stands 1, 5, 6, 9 and 10 may be more immature or pioneer stands, whereas stands 2, 3, 4, 7 and 8 may be more mature or climax stands.It is not implied, however, that these stands are part of the same successional series.
In Table 3 the dominant species of each layer for each category are shown.The figures indicate impor tance as a percentage of the total basal area for the stand, and are determined as shown in the following example: Protorhus longifolia is the dominant canopy layer tree in stand 1, having a basal area of 4 492 of the total 5 552 square inches (see Table 2); 4 492 is 23,9% and 5 552 is 29,5% of 18 792, the total basal area for stand 1.By comparison of Tables 2 and 3, the importance of each species in relation to the layer it dominates and to the stand as a whole can be seen.Also listed in Table 3 are the three leading species of each stand, with a fourth leading species sometimes shown to emphasize certain trends.Only the first three are totalled in each case, however, to keep the figures on an equivalent basis.Note that Cola natalensis in stand 8 with a relative importance of 48,8% is the most dominant species of any stand.Note also the totals for the three leading species range from 38% in stand 9 to 71% in stand 8.

Ordering o f the stands
The index of similarity of Czekanowski, as discussed (1959), was used to compare the similarity of each of the 10 stands with every other stand.The comparison was done on the basis of species present in the stands using the 101 species which were found in the sample.
The formula used is -^-u where a and b are the a + b numbers of species present in each stand, and w is the number common to both stands (Curtis, 1959).
Stands 1 and 8 had the lowest similarity and were, therefore, the end stands.The other stands were arranged according to their similarities to the end stands, and the resulting order is used in Tables 2  and 3 (see Table 4).
Not enough stands were studied for a valid twodimensional or three-dimensional ordination.How ever, it can be seen that the groups of stands which have the highest indices of similarity with one another are those within one forest area.For example, stands 1, 2 and 3 of forest area A have more in common with each other than with any other stands, as do stands 9 and 10 of area B, etc.
The values for the stands within each forest area compared to every other forest area were then averaged, and an average similarity within and between the forests is shown in Table 5. Stands 1, 2 and 3 of area A average 52% similarity with one another, stands 9 and 10 of area B are 61% similar, etc., whereas forest area A averages 40% similarity with area B, etc. Table 6 shows that whereas areas A and B average 40% similarity, and areas C and D average 37% similarity, areas A and B average only 22% similarity with areas C and D.  slopes, compared to (CD), the stands along the coast on dune sand and generally on more gradual northand east-facing slopes (see Table l).

Patterns o f species distributions
Strychnos madagascariensis is one of three leading species in forest areas C (stands 4-6) and D (7-8), and Celtis africana is one of the four leaders in these two areas, except for stand 5 in which it was fifth.Cola natalensis and Cavacoa aurea are dominants in area D (Fig. 4), but are absent from area C. Con versely, Trichilia dregeana and Bequaertiodendron natalense are dominants in area C (Fig. 5), but are absent from area D.
Species patterns are less distinct in areas A (stands 1-3) and B (9-10), although Protorbus longifolia is first or second in each stand, except stand 2 in which it is sixth.Tricalysia capensis, a shrub present in seven of the 10 stands, actually achieved a rank of third leading dominant in stand 1 (Fig. 6).Millettia  We may therefore conclude that the first order of similarity is within one forest area (1, 2, 3) (9, 10) (4, 5, 6) (7, 8).The second order of similarity is (AB), the stands farther inland, altitudinally higher, on rocky soil, and in a topographically more mesic situation with generally fairly steep southwest-facing grandis appears as a dominant in stands 9 and 10, the only stands in which it is ever present.Podocarpus latifolius, a very conspicuous tree in the forest, is present only in stands 1-3, and is a dominant only in stand 2 (Fig. 7).Strychnos madagascariensis is one of the three leading dominants in six of the 10 stands studied and may therefore be considered as the single species most likely to be a dominant in these Coast Forests.Bequaertiodendron natalense is present in eight of the 10 stands, the highest frequency for any species.Only two species appear in all four forest areas how ever: Chaetacme aristata and Tricalysia capensis.
The distribution patterns of other species may be seen in Table 2.The number of species of the various layers as appearing in the two main groups of stands, the inland, rocky areas (A and B) and the coast dunes (C and B), may be summarized as follows: Although there are definite patterns in the distri bution of species in the various stands and forest areas, no definite groups of species with identical patterns are obvious.Jt appears likely that the behaviour indicates a continuum rather than an association interpretation.Further studies covering a wider range of successional conditions are desirable before we can fully understand these distribution patterns.

Soil analysis
From an analysis of the soil samples there were a few trends apparent within the groups of stands.The soils of areas A and B had more coarse sand, more clay, lower pH (4,2 to 5,0), and higher waterretaining capacities, while those of areas C and D had more medium and fine sand, higher pH (6,0 to 7,0), and lower water-retaining capacities.The soils of forest area D had the highest total bases, but the results for the other areas varied with the stands.Area D, likewise, had the highest cation exchange capacity, followed by areas A, C and B. Results concerning specific nutrients varied from stand to stand with no real pattern.If a larger number of stands were studied, it is possible that presence or absence of certain species might be correlated with certain nutrients.

Heterogeneity and homogeneity o f stands
The terms heterogeneity and homogeneity are relative.It is possible to prove that a stand is hetero geneous, but it is unlikely that a stand of any size can be demonstrated to be homogeneous.Consequently, we can speak only of higher or lower degrees of heterogeneity.
(able 7 gives an indication of the relative degree of heterogeneity ot each stand.Sonic stands may be considered less heterogeneous than others.The left hand columns show the decreasing rate at which new species were sampled in the successive plots of a stand.While the figures are similar to a species area curve, ii is realized that chance may have influenced limits to recognize the layers, rather than deter mining each individual subjectively in the field, the time spent in sampling a stand is shortened.Likewise, the relative contribution of the canopy, subcanopy and understorey species to the various layers can be readily assessed and is on a comparable basis for all stands.the results.However, the results are intended to show that there is probably some degree of vegetational unity m each stand.Completely heterogeneous stands would perhaps show a relatively constant rate of encountering new species in successive plots.With this interpretation, stand 7 may be more heterogeneous than the others.
The right hand columns of Table 5 illustrate the frequency with which species appeared in the plots of each stand.Relatively more spccies having high frequencies indicates relatively less heterogeneity.Stands I. 9 and 10 could be considered as the least heterogeneous stands by this method.A high per centage of species occurring 111 one plot only would imply that a stand is relatively more heterogeneous, as is stand 2 by this interpretation.
A relatively large number of species may indicate complexity tor a stand, but docs not necessarily indicate heterogeneity, fo r example, stand I had 40 species, yet seems less heterogeneous than most stands as 55",, of its species have a frequency of (■>()",, or more.Conversely, stand 7 has only 22 spccies yet mav be more heterogeneous than most stands.
( ONC I UNIONS The aim of this paper was to demonstrate the feasibility of applying quantitative phytosociologieal techniques in fairly complex subtropical forest.I rom the re su lts obtained it is apparent that such a quan titative description is feasible and yields useful results. I lie phenomenon of layering, or stratification, is generally considered to be more pronounced in tropical and subtropical than in temperate forests.Although discontinuous layers were not looked for 111 this sttn.lv. the species have inherent growth properties which permit or do not permit them to be come large canopy trees.Stratification, as a structural characteristic of these forests, was described by using specific si/e limits to classify species as to those commonly found 111 the canopy, subcanopy, 0 1 undcrstorev.In the 10 stands, a total of 101 species was found, of which 3X were classified as canopy spccies, 49 as subcanopy species, and 14 as understorey species (i.e., shrubs).By using objective, indirect Basal area in square inches per acre was used as an indication of the relative dominance of the species and of the various layers.Basal area is an objectivelydetermined measure which rellects both the size and the number of trees, and each spccies and layer can be expressed as a percentage of the total basal area of the stand.The relative basal area figures are con sidered to give an approximation of the relative influence on the environment of the various species and layers.
The subtropical forests of the Natal coast were found to be complex, having a large number of species, genera, and families, and this complexity is enhanced by layering (Table 2 and Appendix).The forests were al o found to be fairly heterogeneous (Tables 2 and 3).There arc definite trends in the distribution of species, with certain spccies tending to occur together with other spccies in similar habitats (Table 2).There is, therefore, a pattern in the variations from stand to stand and from area to area as illustrated in the similarity figures (Tables 4. 5 and 6).Only a few species are relatively important in each stand, and many of the same species arc dominant in several stands (Table 3).
The 10 stands studied were in four forest areas.In each case the stands within one forest area had the highest degree of similarity with each other.Two of the forest areas were located on coast dunes and the other two were farther inland, altitudinally higher, and on steeper, rocky slopes.As would be expected, the dune forests were more similar to one another, as were the inland forests to each other.We may, therefore, conclude that the first order of similarity was geo graphical proximity, and that the second order was topographic and soil similarity (Tables 4. 5 and 6).Proforluts longifolia was the dominant canopy species most characteristic of the inland forests, while Strychuos madagascariensis and Celtis africana were most characteristic of the dune forests.There is an indication that the inland and dune forests represent two different communities, and that the five stands in each community represent some sort of successional trends.However, of the 77 species occurring in more than one stand, 43°" of them occurred in both inland and dune stands, and no indication of discontinuous species groups or associations was noted.
An interpretation of the behaviour of the species is that of all the species present in a stand only a small percentage of them is dominant and present throughout the stand, while a large percentage occurs in very small numbers.The relatively large number of individuals appearing only occasionally in the stands leads to the interpretation of the stands as being heterogeneous.However, as seen in Table 3, only two or three species are required to attain more than 50% dominance in eight of the stands, com pared with the total of 20 to 40 species found in each of the 10 stands.Four species are necessary in stand 2 and six species in stand 9 to attain 50% dominance.It is likely that this same situation regarding dominants prevails in temperate zone forests, with two or three species being dominant.However, in temperate forests, only a few other species would constitute the remainder of the stand, compared with these sub tropical forests in which a large number of species constitute the remainder.
The sampling method described in this paper is relatively rapid in the field and requires a minimum of field equipment (a range-finder and a basal area tape).The analytic methods are relatively simple, yet much information is derived.The use of several smaller plots is preferable to one large one in that this avoids basing judgement on one atypical area and allows for variability within a stand.Plots of 1/10 acre are not too large to use and simplify the analytic methods.Circular plots have a smaller perimeterto-area ratio than any polygon with the same area, which helps reduce heterogeneity.By use of a range-.*inder to delimit the plots, the difficulty of staking out quadrats is avoided.Sampling of smaller shrubs and herbs could easily be incorporated into the study by using several small quadrats per plot.With data from a larger number of stands, a statistical treatment of the forest communities of a larger region could be achieved.
The first criterion of stand selection was topo graphical homogeneity, and then visual vegetational homogeneity from a structural standpoint.Sampling and analysis of these stands then followed, and we believe that floristically defined vegetation units have been demonstrated to exist despite some hetero geneity.We therefore conclude that the analysis was justified and that the methods described offer the possibility of a broader comparative study of the relatively complex subtropical forests of Natal.

A CK N O W LED G EM EN TS
This study was conducted while the senior author was visiting lecturer in botany at the University of Natal and visiting biologist with the Natal Parks, Game and Fish Preservation Board.Financial assistance, and the use of office space and facilities, are gratefully acknowledged.Appreciation is expressed by the authors to Mkhize for field assistance and to Hulett's Sugar Co., Natal Estates (Pty) Ltd, and the Natal Parks Board for permission to carry out these studies in areas under their jurisdiction.

A PPE N D IX
A total o f 101 species in 78 genera and 39 families was sampled in the 10 stands and the taxa are listed in systematic order (Phillips, 1951).The 24 species which were omitted from Table 2 are included here, and in parentheses after each of these species is the letter C, S, or U , indicating the layer, and a number referring to the stand in which it occurred.
Pig 2.-Aerial view of Hlogwene Forest, location of stands 4-6, looking south.The Indian Ocean is on the left, and the Tugela River in the foreground.Most dune forests of this type along the coast have been replaced by sugarcane plantations, or, near Durban, by housing development.(Photo: Natal Mercury).

F
shown.The large tree is Podocarpus latifolius, the smooth-barked trees are Drypetes gerrardii and the small trees with fluted trunks are Bequaertio dendron natalense.Note the presence of sandstone rocks fallen from cliff's to the left (see Fig.2) and the cycad, Encephalartos villosus, in the ground layer.

Table 1
for location of stands in areas A to D. The few remaining similar natural forests along the Natal Coast arc located in areas marked X. TA B L E 1 .-Characteristics of the ten stands ot Natal Coast Forest which were sampled had fallen from the cliff, and weathered sandstone was probably the main soil parent material.Stands 9 and 10 were located in Stainbank Nature Reserve, 16 km southwest of Durban (area B on the map), and both stands were on Table Mountain Sandstone.Stands 4 to 6 were in the Hlogwene Forest, which is 80 km north of Durban near the town of Stanger (area C on the map).Tr.e soil parent material was dune sand.This forest is owned and protected by Hulett's Sugar Co., and is pictured in Fig. 3. Stands 7 and 8 were in the Hawaan Forest, 19 km north of Durban (area D on the map).This forest is owned by Natal Estates (Pty) Ltd and is likewise on dune sand.

TABLE 2 .
Number of individuals per acre and basal area per acre, in square inches, for 77 species, by layer, in the 10 stands

TABLE 5 .
-Average percentage similarity of the four forests with one another Relative degrees ot heterogeneity tor the stands 011 the basis of species appearance in the sample plots (see text for