July 11, 1919] 



SCIENCE 



31 



,amine the plant in its wild state, sought to 

 find out its known localities hy consulting the 

 larger American herbaria. He was surprised 

 to find specimens from only two localities, one • 

 in Perry County, Pennsylvania, the other in 

 Sussex County, Delaware. The original lo- 

 iCality assigned to the plant by Michaux in 

 1803, namely, near Winchester, Virginia, is 

 almost certainly a mistake, and its occurrence 

 at the localities in Bath County, Virginia, 

 Greenbrier County, West Virginia, and Polk 

 County, Tennessee, in which it is alleged to 

 have been found, appears not to be substan- 

 tiated by specimens in any American her- 

 barium. 



On July 13, 1918, under the guidance of Mr. 

 Kelsey, the writer visited the Pennsylvania lo- 

 cality. The plant is confined to a single patch 

 extending for a distance of over 400 yards 

 along the slope and shoulder of a timbered 

 west-facing hill. It occupies an area of about 

 eight acres. The soil is a loam of buil-gray 

 color, weathered from an underlying shale of 

 similar color. It has no other supply of water 

 than direct rainfall. Overlying the loam is a 

 layer of upland peat a few inches in thickness, 

 such as commonly characterizes an area of 

 acid-soil dry-land vegetation. 

 , The character of the vegetation is indicated 

 by the following list of its commoner plants. 

 Trees: 



Scarlet oak (^Quercus coooinea), 



WMte pine (Pinus strobus), 



White oak {Quercus alia), 



Chestnut (Castanea dentata), 



Chestnut oak (Quercus montana), 



Bed maple (Acer rubrum), 



Black gum (Nyssa sylvatica), 



Dogwood (Cornus florida). 

 Shrubs : 



Box huckleberry (Gaylussacia irachycera) , 



Laurel (Kalmia latifolia), 



Wintergreen (Gaultheria procumbens) , 



Dry-land blueberry (Vdooinium vacillans), 



Trailing arbutus (Epigaea repens), 



Lowbush blueberry (Vaocinium angustifolium) , 



Juneberry (AmelarKihier canadensis), 



Pink azalea (Azalea nudiflora) , 



Sweet fern (Comptonia peregrina) , 



Pipsissewa (ChimapMla umbellata) , 



Spotted pipsissewa (Chimaphila maculata), 

 Witohhazel (Hamamelis virgvniana). 



There is no indication from the soil, mois- 

 ture conditions, exposure or accompanying 

 vegetation that there are any special conditions 

 on this area, diilerent from thousands of other 

 areas in the Apalachian region, to explain the 

 presence of the box huckleberry in this par- 

 ticular spot. 



Over the whole area the root mat of the box 

 huckleberry is practically continuous. Only 

 one isolated piece was seen outside the main 

 patch, and that was on a steep grassy slope im- 

 mediately north of the main area, where farm 

 cultivation had been begun but later aban- 

 doned. This plant is undoubtedly a piece of 

 the original patch, cut off from the rest by the 

 cultivation but left alive because the cultiva- 

 tion had been discontinued. 



The abrupt termination of the patch, unac- 

 companied by any change in the soil, and the 

 absence of isolated patches were most amazing. 

 In the two hours we spent at the place we 

 sought for the explanation, and I think we 

 have it. But before it is outlined, let me 

 present additional evidence of the complete- 

 ness of the plant's isolation. On the north 

 and much of the east and west sides the patch 

 is bordered by cultivated fields and a road- 

 The natural extension of the patch in these 

 directions is therefore impossible. On the 

 south end, however, the present margin of the 

 patch is located amid natural surroundings. 

 It runs through the timber in a sinuous but 

 definite line which coincides with no topo- 

 graphic or other natural barrier. All along 

 this line the patch is actually progressing and 

 extending by sending out rootstocks which 

 throw up new stems at the end of each year's 

 growth. 



For a distance of 125 yards at the southern 

 end of the west side of the patch, the mat ends 

 abruptly at the bottom of the hill at a natural 

 barrier, a slender woodland streamlet, so 

 low at the time of our visit that in several 

 places no water was flowing over its wet gravel 

 bed. For much of the distance the brooklet 

 has slightly undercut the hill, so that the edge 

 of the root mat hangs suspended a few feet 



