MARCOVITCH

MARCOVITCH described Citrus Coji MARC. in a paper issued in 1921, but it was overlooked on account of the impracticability of accession of the original Russian publication appeared as a local experiment station report of Black Sea region. Critical examination of the original description and later Dutch and English summary published in Java in 1926, soon revealed that this name is nonvalid, by the reason of (1) the lack of definiteness and clearness of essential specific characters that involves, (2) the founding of species upon aggregate types, each one of which being distinct species, like Citrus depressa HAYATA, Citrus tachibana TANAKA, Citrus kinokuni HORT., Citrus tangerina HORT., Citrus erythrosa HORT., et., (3) the failure of distinguishing the description of the type of the species, thereby causing impossibility to reserve the name, and (4) the impossibility of applying the name to the Koji mikan of Japan by the antagonistic diagnosis such as wild nature, bitterness of the pulp, and the like. Since these points are enough to become the permanent source of error, and since the Koji mikan is a definite horticultural species and must have a name, the author reached to the conclusion that the name Citrus Coji should be proscribed, and a better name should be chosen for Koji. Citrus koosi SIED. in sched. is based on imperfect herbarium materials, and is not to be used on expectation of future confusion. Citrus koozi HORT., offered in a paper of the author read before the Third Pan-Pacific Science Congress (based upon Citrus nobilis subsp. Suntara var. Koozi TANAKA. nom. seminud. published in 1912) is withdrawn to avoid further mistake. Citrus nobilis var. Koji found in popular Japanese books is also rejected by being applied to aggregate species lacking the description of the type. From these unfortunate circumstances, the author is compelled to create a new horticultural species Citrus leiocarpa HORT. nov. based upon Tsuchi Koji or Mar Koji of the Japanese gardens, which covers subordinate horticultural varieties Hira Koji (Kin Koji), Suruga Yuko, and Sagami Yuko (Fukure mikan). Citrus leiocarpa is a good horticultural species, being a small thornless tree, dencely convered by broad leaves of medium-small size and acuminate apex. Flower-buds are broad, and their cross section is angular, the calyx lobe of which being much rounded with deep and sharp cut and glabrous margin. Fruits are small, depressed-globose, tumid, much polished and yellewish-orange in color, with smooth, minute oil cell dots. Segments are 8-11, regular and moderately thick-walled, leaving small space at the central column; pulp, is deep yellow, narrowly pararell-netted, with juice of mild subacid flavor. Pulp vesicles are long-fusoid, separable and not delicate. Seeds are rather large, ellipsoid, green, not beaked, smooth, containing green polyembryos; tegmen at charaza is marked rosy-brown. Its closest wild relative is Citrus depressa HAYATA, never occurring north of Sanbok (Amami Oshima), which differs from it by having broadly pointed leaves, reddish rind, finely round-netted pectiniferous pulp vesicles, large seeds of brownish coat, and dark-brown charaza part of tegmen. Another wild relative, Citrus tachibana TANAKA of Japan proper, has soft fruit of decidedly small outline, truncate, broadly depressed ends, baggy rind, fewer number of segments, melting acid pulp, short-fusoid extremely delicate pulp vesicles, smaller seeds, and very light-colored charaza part of tegmen. (Joint contributions from Horticultural Institute, Kyushu Imperial University, No. 12 and Phytotechnical Institute, Miyazaki College of Agriculture, No. 20.)