Experiments on the cultivation of dandelion for salad use. I. Study of cultivation methods and their influence on yield and sensory quality.

Cultivation of mild-tasting dandelion types for salad use has been studied as one part of an investigation aimed at evaluation the possible uses of dandelion. Eleven Finnish agamospecies were compared with two middle-European varieties, ’Vollherziger, verbesserter’ and ’Vert de Montmagny’ in regard to preferred methods of cultivation as well as to yield level and quality properties. Further, suitability and effects of the culturing methods were studied, the main variables being level of fertilization and use of various coverings to bleach the rosettes. Of the cultivation properties, particularly the earliness, flowering, winter hardiness, vigour of growth and resistance against diseases and pests were considered. The yields in spring varied between about 50 and 150kg/are. When growth was in light the bitterness of all the agamospecies and varieties studied was only slight and when effective covering was used bitterness was insignificant. The sensory quality varied somewhat depending on the agamospecies or variety and method of cultivation, being at best very good and at worst satisfactory. In the conditions studied, the higher level of fertilization was too high. The single black film, the black fibre cloth and thick bark humus layer were all suitable coverings. Though they decreased the yield (at least when calculated as dry matter), the coverings improved the taste and texture of the rosettes. With film and fibre cloth, rosettes were somewhat more greenish and yields better than with bark humus. The humus also made the rosettes more succulent and crisp, so that the texture was improved. This method is laborious, however, and better suited for home gardening than commercial cultivation. Of the agamospecies and varieties studied, the best for cultivation in light were Taraxacum hemicyclum, T. lanyphyllum, ’Vollherziger verbesserter’ and ’Vert de Montmagny. For cultivation under covering the best were ’Vert de Montmagny’, T. hemicyclum and T. pallidipes. Index words: dandelion, salad greens, cultivation methods, agamospecies, fertilization, bleaching, yield, sensory quality, bitterness 9 JOURNAL OF AGRICULTURAL SCIENCE IN FINLAND


Introduction
The possibility of utilizing the dandelion has been studied since 1977 by a group of researchers, originally from the Food Research Laboratory of the Technical Research Centre of Finland and the Department of Botany at the University of Turku, and later further from the Department of Horticulture at the University of Helsinki. The aim has been to investigate the various possible uses of the dandelion, which flourishes throughout Finland. At the start of the study, 157 samples of Finnish dandelions were removed to the Botanical Garden of the University of Turku. These samples were identified and found to represent 69 different agamospecies. The botanical, cytological and genetic properties of the agamospecies were studied in Turku by Rousi and his co-workers (1984, in press), while the degree of bitterness was investigated by sensory methods at the Food Research Laboratory. The bitterness is an interesting property: if dandelion is used for taste-giving purposes e.g. in drinks, strong bitterness is an advantage, whereas if it is used as a salad ingredient, the bitterness should be minimal. The bitterness properties will be discussed in a forthcoming paper.
The possibility of cultivating the dandelion commercially has been studied since 1980. On the basis of the bitterness tests, those agamospecies most promising for cultivation as a salad green were selected for further investigation. The most important properties for such use are mild taste, and vigorous growth as a basis for sufficient yields. Winter-hardiness, resistance against diseases and pests, general sensory properties and nutritive value are important as well, but about these was little knowledge when the experiments were begun.
Elsewhere in the world, particularly in the Unites States and France, dandelion has been cultivated for salad use, and information is available on methods of cultivation (see the review by Kanon 1982). The dandelion has apparently not been cultivated for food in Finland, though it was once experimentally grown for its rubber content (Suomela 1950). One important cultivation method for salad dandelion involves bleaching the rosettes. In this respect it was of interest to experiment with some new methods of horticulture, such as covering the plants with black polyethylene film, black fibre cloth, or layers of bark humus, instead of using the rather laborious methods suggested in the literature. The cultivation experiments were designed and carried out at the Department of Horticulture, and the Food Research Laboratory performed the sensory analyses and evaluations of bitterness. These results are reported in this paper, and the nutritive value and intrinsic quality properties such as nitrate content are reported in part II which follows (Kuusi, Härdh and Kanon 1984 pp. 23-31).
Full details on the cultivation experiments can be found in the pro gradu work of Kanon (1982). Experiments concerning the winter-forcing of salad dandelion will be reported at a later date.

Material and methods
Material. The material consisted of 11 Finnish agamospecies, selected from the 69 agamospecies primarily on the basis of their mild taste ( = low degree of bitterness) and vigorous growth. For comparison, two foreign cultivars were included in the experiments: 'Vollherziger, verbesserter' Mauser of Swiss origin, and 'Vert de Montmagny' Vilmorin of French origin. The seeds of the Finnish agamospecies were collected in 1978 and 1979 from the experimental cultures in the Botanical Garden of the University of Turku, and stored at + 5°C. The seeds of the foreign cultivars were obtained commercially. The agamospecies and varieties are listed in the tables.

Design of the cultivation experiments:
Two main experiments were performed: 1) comparison of the different agamospecies and varieties and evaluation of bleaching with black film, and 2) comparison of two levels of fertilization and of different methods of covering. The experimental method was that of split plots with four replications.
Cultivation methods: The seedlings were grown in a glasshouse, where mean temperature was 20°C and humidity 71 % (8 o cl.). Seeds were sown in April in sxB cm peat pots (FP 620), filled with light, medium porosity, medium fertilized peat (St-400 B 2), two seeds per pot. The seeds were covered with a 0.5 cm layer of sand and protected against evaporation with colourless film until after germination. Later, the less advanced seedling was removed from each pot. In the first experiment comparing the agamospecies and varieties the glasshouse cultivation was continued for 6 Vi weeks, in the second, fertilization and covering experiment for nearly 6 weeks. In the second experiment the variety used throughout was 'Vert de Montmagny'. Planting outdoors took place the end of May. Planting distance was 20 cm and distance between the rows 50 cm.
Care of the cultures included mechanical weeding and weeding by hand. The flower stalks and buds were removed every week. During long periods of dry weather irrigation was applied, four times approximately 15 mm each time.
Fertilization and use of coverings. The experimental field was inclined slightly toward the northwest, the soil type was medium-humus fine sand. The area had been ploughed the previous autumn. Before basic fertilization the nutrient level was according to analysis (by Viljavuuspalvelu Oy) the following: for Gardens Y I* was applied, in the second experiment either 8 or 16 kg/are of the same mixture. In the second experiment an additional 1 or 2 kg/are calcium nitrate (Ca(NO 3 ) 2 , 15.5°?o N) was applied in April before laying of the covering. The nutrient status of the soil was checked in the second spring before the beginning of growth and after harvesting.
The coverings used were bark humus (bark humus of Kirkniemi, lime added and basic fertilizing done), black polyethylene film (thickness 0.05 mm), and black slitted »Fibertex» cloth. The bark humus was applied 1) before planting of the seedlings outdoors, as a layer 2.5-cm thick, or 2) after the plants had withered in November, as a 5-cm thick winter covering, or 3) in addition to 2), at the end of April as a further layer of 5 cm. In this last case the covering was intended for bleaching of the rosettes. The other coverings to bleach the rosettes were similarly applied at the end of April, when the uppermost 10 cm of the surface of the earth was unfrozen and the dandelions had begun to grow. The black film was supported by arches ca. 50-cm high in the first experiment, while in the second experiment the covering had no support. The fibre cloth was similarly unsupported. The edges of both the film and fibre cloth were anchored with earth.
In one part of the covering experiment the black film was applied double; when single, the bleaching was insufficient. When the temperature rose too high the film and fibre cloth coverings were opened to allow ventilation.
Observations in the field during the Basic fertilization was carried out before planting the seedlings outdoors. In the first experiment 10 kg/are of Fertilizer Mixture * composition, %: Main nutrients N-P-K = 10-4-17 (P 2 O s , K 2 (D), B 0.15, Cu 0.4, Mn 0.7, Mg 2.5, Mo 0.02, Zn 0.03, Fe 0.1 growing season. The viability of the plants was checked after transfer to the field in June and at the end of the growth season in September, the winter hardiness at harvest in the spring.
The start of flowering was noted during the first summer and the flower stalks were counted every week. In spring the rosettes were inspected for flower buds. The earliness of withering was checked at the end of the season, and the earliness of sprouting at the end of April, where this was not prevented by the coverings. Numerical scales were used to register the differences. The diseases and pests occurring were identified at the Departments of Plant Pathology and Agricultural and Forest Zoology of the University of Helsinki. The damage was followed and the degree assessed using a numerical scale (0-4) weekly or every other week.
In the first experiment harvesting took place the second spring between May 18 and 22, in the second experiment on May 19 and 20. The rosettes were cut ca 1 cm below ground level to keep the rosettes intact.
Withered leaves were removed and the rosettes stored at + I-2°C, wrapped in plastic pouches.
Yields were determined by weighing the harvested and cleaned rosettes and calculated as kg/are. Weighing was carried out in spring and in August of the first year.
The sensory evaluation was done according to established principles (see e.g. Ame was used (see Paulus et ai. 1969) with separate points given for appearance and colour, texture, smell and taste. Appearance, colour and smell were evaluated from whole rosettes, texture and taste from leaves that had been shredded without removal of the vein and served as such. (Fig. 1 shows the evaluation form.) Bitterness was evaluated from extracts prepared immediately after harvesting from the replicate leaf samples. Thus, 18 g leaves was homogenized in 100 ml distilled water with a Bamix for 2 min., after which the homogenate was boiled for 2 min., cooled, filtered and diluted to 100 ml with distilled water. The degree of bitterness was evaluated by 5 trained panelists using a numerical scale between 0 and 6, where bitterness value 3 ( = medium) was fixed with a quinine standard of strength 0.0015 % quinine in distilled water. (Fig. 2 shows the evaluation form.)

Results
Observations in the field. The earliness of growth in the second year was evaluated at the end of April using a numerical scale o -3 . 0 -3. (Fig. 3). The agamospecies Taraxacum alatum, T. ekmanii and T. hjeltii were the earliest, with the foreign varieties closely behind. A similar order was observed at harvest. Further, dandelions may then be transferred to undesired places. The earliest to flower was T. hjeltii, followed by T. alatum, T. amplum and T. aculeatum in July-August. The others began to flower at the end of September of not at all. Figure 4 shows the trend for flowering during the first year. In the following spring the rosettes were harvested before flowering, but in all cases flower buds were present. Use of covering to bleach the rosettes delayed the flowering for one or two weeks.
The extent of withering was checked at the end of the growing season, 6 Nov. 1980 (Fig.   5). By that time, the early agamospecies T.
hjeltii was already fully withered. The figure makes clear the considerable variation in the withering of the different agamospecies and varieties. The winter hardiness was checked after the winter and it was noted that, in all, 6.7°Jo of the plants perished during the winter. In this respect, however, there were no systematic differences between the agamospecies and varieties. The diseases were most conspicuous from mid-August until the end of September, but in spring they did not harm the rosettes to be harvested. Figure   In addition, damage by voles to leaves and roots was noticed in late autumn, and 16 plants were destroyed (1.9 %). The yield results are shown in Table 1. Average yields of the foreign varieties were mostly somewhat higher than those of the Finnish agamospecies, though individual yields of some of the latter were at the same level. The fresh weights were higher in late summer than in spring in many cases. However, salad dandelion is typically harvested in spring, and thus the higher yield in late summer would ordinarily not be utilized. The influence of darkness is clear-cut: yields tend to be much lower in darkness.
The higher level of fertilization did not increase the yield. Of the coverings, the bark humus systematically decreased the yield, and the black film and fibre cloth increased it.
The results of the sensory evaluations in Table 2 show slightly higher averages for the foreign varieties than the domestic agamospecies. For cultivations in light, the scores of taste and texture are generally lower than the scores for colour, appearance and smell. Bleaching systematically improved the scores, in particular taste and texture, and to some extent colour. Some of the Finnish agamospecies have scores as high as the foreign varieties, although the latter show a clearer superiority when bleached. The scores for the two fertilization levels are very similar: fertilization had no influence on the sensory properties. In contrast,  covering improved the sensory properties if definite bleaching was obtained. Of the bark humus coverings only the 10cm covering was effective in this way.
The results of the bitterness tests are shown in Table 3. In this respect the different agamospecies and varieties varied markedly, even though all had been selected for their mild taste. The range of values was widest in the late summer when the bitterness was also greatest. The foreign varieties tend to be more bitter than the Finnish agamospecies. However, the bitterness is low in all: on the scale used here, 3 represents medium bitterness, and in the spring dandelions all scores were below this, varying between 1.3 and 3.0 when cultured in light and between 0.8 and 2.0 when cultured in darkness. Salad dandelion is not ordinarily harvested in late summer when the bitterness is greater. The level of fertilization has no clear-cut influence on the degree of bitterness, but covering clearly decreased the bitterness if bleaching was obtained. A thin layer of bark humus was not sufficient for this effect.

Discussion
Among the properties studied here, the most important for the future use of the dandelion as a salad green are yield, sensory quality and degree of bitterness. Certain properties of cultivation will have to be taken into account as well. In the following the results are evaluated from the point of view of such use.
The yields of the different agamospecies and varieties varied widely, with Taraxacum hemicyclum and 'Vert de Montmagny' grown in light performing best. The level of yield corresponded to that obtained in East Germany (see Hahn & Mildnerl962), but was lower than the yield considered good in the United States (Sackett 1975). The weights of the rosettes were greater in late summer of the first year than at harvesttime, the following spring, but the sensory quality and nutritive value were essentially better in spring. It may be added that salad dandelion is more reasonably harvested in the spring, when other salad greens are unavailable or expensive.
Although the dandelions were harvested by cutting below the upper end of the root, all except T. hjeltii regenerated new rosettes about one month after the harvesting, which offers the possibility of increasing the yield. The features of regeneration have been studied further in experiments to be reported separately.
The sensory quality was in general either satisfactory or good, and the best agamospecies and varieties, such as T. hemicyclum and 'Vollherziger, verbesserter' were good even without bleaching. In the spring harvest the differences in the sensory quality were only slight.
The bitterness varied somewhat between the different agamospecies and varieties, being clearly less in spring than in late summer, in agreement with the results of Huttunen (1981). There was no correlation between the degree of bitterness and the intensity of the green colour of the leaves, typical for agamospecies, nor between bitterness and the presence or absence of red colour of the middle vein or the type of lobes in the leaves (see also Huttunen 1981). Such characteristics can therefore offer no guidance when wild dandelions are being collected. In the dandelions studied here the bitterness was consistently slight, as intended.
Of the cultivation properties, earliness would be an advantage. In the spring of the second year, growth was sufficient from the middle of May (though that year the spring season was two weeks later than usual). The earliest agamospecies was T. hjeltii, which comes from northern Finland. The difference in time of harvesting was at most one week, the latest species being T. ancistrolobum, T. tanyphyllum and T. pallidipes. Hot weather may well diminish the differences in earliness, and thus this property may not be very important.
The period of proper harvesting is comparatively short, since immediately the rosette of leaves has reached the desired size, the flower stalks begin to lengthen and bitterness rapidly increases. This period could be somewhat prolonged by growing several agamospecies or varieties differing in earliness.
The differences in flowering were conspicuous: during the first growing season only a good third of the samples flowered. Where the flowers are not needed, it would be best to select species that do not flower during the first summer.
No definite differences were noticed in winter hardiness: all species studied were sufficiently resistant. The winter of the experiment, it may be added, was less favourable than average.
Pests, except perhaps voles, are not a threat to dandelion cultivation. Three main fungus diseases were observed, of which par-ticularly powdery mildew could become harmful. Rust and leaf spot were also of importance, although they occurred mostly in late autumn. Grey mould was insignificant. The diseases do not influence the quality of the rosettes to be harvested in spring, but as general factors weakening the plants they may decrease the yield.
Of the cultivation methods, a higher level of fertilization did not increase the yield in the conditions studied. Only when bark humus was applied during the growing season, causing a need for more fertilizer, did the higher level of fertilization give better yield than the lower. The results do not exclude the possibility that the optimal level may be higher or lower than the lower level used here.
Bleaching by covering with black film significantly decreased the yield in the first experiment. In the second experiment, where the different coverings were compared, the yields for plants under film and fibre cloth were greater than for uncovered plants, while covering with bark humus diminished the yield. The results may be connected with the temperature effects of the different coverings. However, when the yield was calculated as dry matter, the yield was always lower when coverings were used, in agreement with earlier results (Wassink 1965).
Use of coverings delayed the beginning of flowering, but the harvest period was not thereby longer, since under film and fibre cloth the rosettes may become too long, and when bark humus covering is used, harvesting must be done before the leaves penetrate the layer.
Bleaching virtually eliminated all bitterness and thereby improved the sensory quality, particularly taste. But also texture was improved: the bleached rosettes were succulent and crisp, and particularly when bark humus covering was used, the base of the petiole was pleasingly fleshy. Certainly the 10-cm layer of bark humus was not sufficient; it would be better to apply at least a 20-cm thickness, given the length of the leaves and the fact that in darkness the rosettes assume a more vertical position than when growing in light.
Comparison of the different methods of covering showed that the double black film was too tight, causing damage to the rosettes. If it was applied single, the bleaching effectiviness was somewhat less than that of the fibre cloth. Where bark humus was used growth was a little delayed, because the covering had to be laid in spring on the cold, still frozen earth, which then warmed up more slowly. Further, the bark humus covering caused a decrease in yield.
The decision to cover, as well as the type of covering, depends on the kind of crop desired. Unbleached the yield is greater, rosettes are pale green, and there is slightly more bitterness. Fibre cloth gives a crop of somewhat better quality, but it is more expensive than black film. A thick layer of bark humus gives the best texture of the rosettes, but application is laborious and the method is thus more suitable for home gardening than commercial cultivation. In addition, with bark humus covering the fertilizer level must be increased.
In comparing the merits of the different agamospecies and varieties, each of the properties studied must be taken into account and its importance weighed. If yield, sensory properties and mild taste, as well as resistance against diseases and pests are considered to be most important, and only half as much weight is placed on earliness, flowering properties, ability of regeneration and amounts of vitamin C and soluble solids, then the most suitable lines for culture without bleaching are T. hemicyclum and T. tanyphyllum, and the two foreign cultivars 'Vollherziger verbesserter' and 'Vert de Montmagny'. With bleaching the most suitable are 'Vert de Montmagny', T. hemicyclum and T. pallidipes. Of the Finnish agamospecies the most suitable is thus T. hemicyclum, where the growth of rosettes is vigorous, the yield high and taste mild, and where the sensory properties are as good 20 unbleached as bleached, and the yield is not significantly lowered by bleaching. This agamospecies is early in growth and does not form flowers at all during the first summer. As disadvantages, a medium susceptibility to both Sphaerotheca (powdery mildew) and Puccinia (rust) as well as also Ramularia (leaf spot) must be mentioned.