Bracken is an indicator of medium to poor forest site quality in Michigan, as observed for a long time by foresters and soil surveyors alike. Its low-site indicator value has probably deferred the study of its ecology in Michigan. The primary purpose of this reseach was to quantify the relationships which exist between bracken size and its site. The secondary purpose was to evaluate bracken as a soil-site indicator for pure stands of aspens, oaks, and jack pine on sandy outwash soils in northern Lower Michigan. """ To achieve these goals, sample areas were located in nine northern Lover Michigan counties which are covered by large deposits of outwash sand: Cheboygan, Crawford, Emmet, Grand Traverse, Kalkaska, Lake, Montmorency, Oscoda, and Otsego. The samples were designed to yield information about forest cover, soils, and bracken growth. During the fall of 1961 and the following year, a total of 352 bracken milacres were measured in association with 86 forest cover and soil profile samples.""" The primary task, of establishing relationships between bracken frond size, forest cover types, and Podzol profile development gave the following major results: 1. Both bracken size and forest growth improve as podzolization proceeds in Michigan outwash sands. For example, averaging all forest covers, a bracken frond on a Rubicon sand will be about 30 cm long. The average bracken frond length was found to be 17 cm shorter on Rubicon sand (minimal Podzol) than on a Kalkaska sand (medial Podzol), and 17 cm longer on Rubicon sand than on Grayling sand ( Podzol absent). On Saugatuck sand (marginal Podzol) the average length may be expected to be 34 cm greater than on Rubicon sand. I did not collect sufficient data to quantify statistically the relationships between the various forest cover types and soil profile development. The fact that site index for forest trees improves with increasing podzolization in Michigan has been recognized previously however (U.S.D.A., 1961a). The positive correlation between bracken frond growth and soil profile development is dependent on several soil characteristics. As the B2 horizon develops in a sandy Podzol profile, its nutritional status and its water holding characteristics improve. These facts are important enough to counteract the possible deleterious influences of the progressively more and more leached A2 horizon. The increasing silt, clay, and VFS accumulation in the B2 horizon as podzolization progresses is a measure of A2 leaching time and of the leaching efficiency of different litter types supplied by different forest cover species. 2. Forest cover types exert a major control over bracken frond size. If all the bracken size data are stratified by cover type, regardless of soil profile development, it appears that open areas support bracken fronds averaging 63 cm in length; oak cover supports fronds 76 cm long, pine fronds 93 cm long, and aspens, fronds 107 cm long. When both forest type and Podzol profile development are considered together, the soil development contributes 16 cm between Grayling and Rubicon sands, 16 cm between Rubicon and Kalkaska sands, and 16 cm between Kalkaska and Saugatuck sands. Forest cover contributes 10 cm between open areas and oak cover, 10 cm between oak and pine covers, and 10 cm between pine and aspen covers. An average bracken leaf on Grayling sand in the open will be about 54 cm long. On a Saugatuck sand under aspen cover, however, it would be 48 cm longer because of the soil change and 30 cm longer because of the forest cover change, a total of 78 cm. The natural average range in bracken size between 54 and 132 cm (54 plus 78 cm) can be accounted for by forest and soil influences on the specific sites under consideration in this thesis. The smallest frond measured for this study was 35 cm and the largest was 159 cm long. The added variation between the average range values and maximum-minimum values is attributable to a host of different influences some of which have been studied here. There are a number of reasons why forest cover influences are i portant in controlling bracken frond size. Observations on the phenology of bracken crozier emergence in the spring revealed that the protection against frost differs under various cover types. Under aspen cover, bracken croziers begin to emerge from the soil in early May in northern Lower Michigan, under jack pine in mid May, under oak about May 20, and in open areas not till the end of May or beginning of June. The forest influences per se is therefore one of cortrol over crozier emergence. The later the croziers emerge the smaller the final size leaves. For each week after May 5 that a crozier emerges it will be 5 cm shorter than if it had emerged a week earlier. This holds true for all the forest cover types studied here. The litter supplied by different tree species also differs in its ability to protect the soil from frost penetration. It probably also plays an important role in speed of soil warming in the spring. When oak litter was removed from sampling areas, avarage frond sizes were reduced 37 per cent. Aspen litter removal resulted in a 20 per cent leaf size reduction, and pine litter a 6 per cent reduction over the same time period. Bracken fronds were not materially affected by doubling the natural litter amounts under these forest cover types. """ 3. Forest shade influences the bracken frond shape and growth habit. Total frond height/stipe height ratios exceed 2.2 for open grown fronds but rarely exceed 1.7 for shaded leaves. Open grown fronds have shorter stipes and considerable segment overlapping in the blade while shaded leaves have longer stipes and little or no segment overlapping. Forest shade also influences the number of stomates per unit of ventral leaf surface and the number of subsidiary cells per stomate. The more light that reaches the leaf surface the more stomates that occur per unit of leaf area and the more subsidiary cells area associated with each stomate.""" """ 4. In order to separate clones of bracken in the field, several criteria must be used together. There is considerable variation in most parameters even among leaves from a single rhizome system. The parameters which I found most useful are: blade length/primary pinna ratios, leaf segment sizes, blade/stipe ratios, and numbers of primary, secondary, and tertiary blade divisions per frond. Even when using such established criteria, the separating of clones in the field is difficult because of this large within-plant variation.""" 5. When I attempted to use bracken as a site indicator for forest trees (oaks, aspens, and jack pine) it became apparent that soil series ( Podzol profile development and forest cover per se could not explain all the variation which was observed in bracken frond size on the same sites. It was evident that bracken is sensitive to frost damage during the winter, spring, and fall, impairing bracken population size, vigour, and leaf shapes. These sources of variation, in addition to local soil variations, microclimate variations between sites, genetical differences in bracken clones, and insect and pathogen damages, change from site to site. The age of the plant section from which the frond arises also affects that frond’s size. In spite of all these sources of variation, however, bracken can be used to classify forest site in northern Lower Michigan sandy soils providing care is taken to avoid obvious frost pocket areas or areas where bracken can be seen to have suffered from environmental influences other than those exerted by the soil and the forest cover.