1University of Derby, School of Environmental & Applied Sciences, Kedleston Road, Derby DE22 1GB, UK; 2ADAS Arthur Rickwood, Mepal, Ely, Cambs. CB6 2BA, UK

Background and objectives
Ever since the 19th century, when large flowered varieties of Clematis (Ranunculaceae) were first introduced on an extensive scale, the cultivation of this ornamental climbing plant has been troubled by a destructive disease, generally known as 'clematis wilt'. The disease is caused by the imperfect fungus Phoma clematidina, but knowledge of the biology of this specific, facultative pathogen is limited, which hampers disease prevention and control. Apart from wilt of above ground parts, symptoms also include leaf spots and stem rot. Even though stem infection of clematis often takes place at soil level and other Phoma spp. are known as soilborne pathogens, root infection by P. clematidina has never been reported. However, an unidentified Phoma sp., closely resembling P. clematidina, was described by Arthur [1] from the roots of wilted clematis plants in the USA. Since root infection could be an important factor in the epidemiology of clematis wilt, the existence of such infections in practice and the pathogenicity of P. clematidina to the roots of clematis were investigated.

Material and methods
Wilted container-grown clematis plants were collected from four nurseries in the UK and pieces of root tissue were surface sterilised in 70% ethanol and plated out on V8-agar. Colonies of P. clematidina on culture plates were distinguished by growth rate, formation of pycnidia exuding pink spore mass and the presence of multi-cellular chlamydospores. To test the pathogenicity of the fungus to healthy roots, six months old clematis plants cv. 'Henryi' were placed bare-rooted in glass flasks filled with water such that only the lower part of the root system was submerged. Ten flasks were inoculated with agar plugs bearing pycnidia and mycelium of P. clematidina and ten with sterile agar plugs. The plants were placed in an unshaded glasshouse during summer and were checked regularly for disease symptoms. After 100 days, roots of all plants were plated out on V8-agar to test for the presence of P. clematidina.

Results and conclusions
Of 33 clematis plants examined, four (12%) had small black root lesions from which P. clematidina was isolated. These plants came from two nurseries and had all been grown in a very moist substrate. In three of these plants, P. clematidina was also detected in above ground parts, such as leaves or stems. With the artificially infected plants, small black lesions started to form on roots after about three weeks, especially at the interface of water and air. The lesions quickly expanded and eventually caused whole roots to rot. Pycnidia exuding a plnk coloured spore mass were formed on many of the lesions. Within 100 days, all ten inoculated plants showed root rot symptoms to some extent and P. clematidina was recovered from all of them. Seven plants wilted, the first ones 35 days after inoculation, but this may have been partly due to hot weather conditions since four control plants also wilted. However, none of the latter showed signs of root rot or yielded the fungus when plated out. In the plants collected from the nurseries, root lesions were not as extensive as found on most of the artificially inoculated plants. The fact that natural infections are more limited suggests that they are less likely to cause wilting, especially in large plants with extensive root systems. Root infections in unwilted plants will usually pass unnoticed on nurseries and are, in effect, latent. Plants with root infections are handled and sold as if healthy and such plants may be important in the spread of the fungus to other nurseries, plant collections and private gardens.

Funding of this work by the Horticultural Development Council (HDC) is gratefully acknowledged.

1. Arthur, JC, 1885. New York State Agricultural Experiment Station Third Annual Report (1884). pp. 383-385.