Index | Search | Home | Table of Contents

link to pdf file

Morales, M. and J. Janick. 2002. Arugula: A promising specialty leaf vegetable. p. 418–423. In: J. Janick and A. Whipkey (eds.), Trends in new crops and new uses. ASHS Press, Alexandria, VA.

Arugula: A Promising Specialty Leaf Vegetable

Mario Morales and Jules Janick


Arugula is a collective name for a number of species of the Brassicaceae with pungent leaves, but principally Eruca sativa Mill. This species has been known since antiquity and is listed in the Greek herbal of Dioscorides (Materia Medica) written in the first century (Fig. 1) as well as the English herbal of John Gerard (1597). Arugula is a low-growing annual with dull-green, deeply-cut, compound leaves which have a distinct spicy-pungent flavor (Palada and Crossman 1999). Like other Brassicas, it contains glucosides such as allyl sulphonocyanate while the seed oil contains erucic acid (Nuez and Hernandez 1994). In various Mediterranean countries it is cultivated as a salad green or cooked vegetable and is also grown in Asia as an oilseed crop. Three species are used for human consumption: E. sativa and a wild type E. sativa vesicaria (L.) Cav., both annuals; Diplotaxis tenuifolia (L.) DC, a perennial species, and D. muralis (L.) DC, a polyploid perennial (Pignone 1997). Arugula is also known as rocket (English), roquette (French), rucola or rughetta (Italian).

170. Euzomon. Eruca sativa Rocket

[the Romans call it Eruca, ye Aegyptians Ethrekicen, the Africans Asuric] This being eaten raw in any great quantitie doth provoke Venery, and the seed of it also doth work ye like effect, being vreticall and digestiue, & good for ye belly. They doe also vse the seed of it in making of sawces, which that it may endure the longer, hauing macerated it first in vinegar, or milke, making it into Trochiscks, they afterward lay it vp in store.

There also grows a wild Eruca, especially in Iberia towards ye west, whose seed the men there doe vse instead of Mustard. It is more diureticall, & farre sharper then the Satiue.

Fig. 1. Arugula is an ancient crop referred to by Dioscorides in the first century (Gunther 1933). English translation is by John Goodyear, 1655.

Arugula is widely consumed in Italy where its pungent qualities are appreciated, either consumed alone as a green, as part of a salad mix, as a cooked green, and now very popular as a pizza topping. The wild form with thin dentated leaves is increasing in popularity. Arugula has appeared in US markets from California production and can be found in select supermarkets as a specialty green and it is often found in farmer’s markets as part of a mesclun mix.

Arugula is a fast growing, cool-season crop and flowers under long days and high temperature. This crop appears well adapted to the Midwest and offers the possibility of a specialty green that could be available over a long period with season extension techniques as it is adapted to greenhouse culture. Our trials indicate it can be harvested after 20 to 27 days and then sequentially harvested from regrowth. European interest in promoting this underutilized crop has prompted the formation of the Rocket Genetic Resources Network, a project of the International Plant Genetic Resources Institute (IPGRI). The proceedings of a 1996 workshop (Rocket: A Mediterranean crop for the world) contains a number of papers on its culture and biology (Padulosi and Pignone 1997).

This study was undertaken to investigate the potential of arugula as a new crop for Indiana and the US Midwest. Our experience with this crop, both horticulturally and organoleptically as a salad green, has embolden us to become champions of this species despite the fact that it is not well known or appreciated in this area. The main obstacle to this crop appeared to be early bolting and this study was originally undertaken to investigate the possibility of selecting late-bolting material as well as to explore the adaptability of this species for Midwestern conditions. We now believe that sequential planting combined with repeated harvest may be a way to manage this crop despite high bolting under high temperatures.


1999 Season

Seed of E. sativa (‘Astro’ and “arugula”) and the wild D. tenuifolia (‘Sylvetta’) was obtained from Johnny’s Selected Seeds, Albion, Maine and planted in 128-cell trays (3 seeds per cell and thinned to a single seedling) in the greenhouses of the Department of Horticulture, Purdue University on June 17, 1999. (‘Astro’ is a selection made by Johnny’s that has large leaves, with less dentation that the “arugula” type.) There were four trays per entry producing an original population of 512 plants per entry. Plants were grown under continuous illumination with incandescent lights (110 to 140 mmol m-2 s-1) during the duration of the experiment. Each tray was placed on a plastic flat which could maintain fertigation water to prevent plants from drying out in spite of the small amount of soil. Plants started flowering 30 days after planting and flowering plants were removed daily when flowers could be distinguished. In mid-August, 18 ‘Astro’, 15 “arugula” and 9 ‘Sylvetta’ plants, which had not yet flowered, were transplanted into 8-L containers and left to flower. However, seed setting was poor (apparently due to self incompatibility) and only a few seeds from 5 ‘Astro’, 4 “arugula” and a single ‘Sylvetta’ plants were collected. The wild type ‘Sylvetta’, with smaller and more deeply parted leaves and smaller seeds, had very poor seed setting and was eliminated from the experiment.

2000 Season

Seed from 5 ‘Astro’ and 4 “arugula” late-flowering selections and the original ‘Astro’ cultivar were planted in trays in the greenhouse on April 10, 2000 and subsequently moved to a plastic greenhouse on May 1; there were a total of 504 plants from the 9 selections and 504 of the original ‘Astro’ cultivar. Visible flowers appeared by May 8 and flowering plants were discarded daily during the next 4 weeks. Thus, two months after planting the trays, 46 plants (9%) were eliminated from the late-flowering selections and 178 plants (35%) were eliminated from the original ‘Astro’. This difference reflects the effectiveness of the initial selection. Non-flowering plants were hardened outdoors and then transplanted into raised beds covered with plastic much and irrigation tape underneath on June 8. Flowering plants were pulled out and discarded up to Aug. 15 at the same time that outstanding late-flowering plants were being flagged. Seed of the most promising late-flowering plants was collected from Sept. 27 through Oct. 3 and then cleaned and stored at 4°C. There were a total of 64 late-flowering selections.

2001 Season

The 64 late-flowering selections (LFS) from the previous season and 3 checks (‘Astro’, “arugula,” and “cultivated arugula” from the Italian commercial source Florsilva) were planted on April 17 in the greenhouse under natural light and seedlings field transplanted into raised beds covered with plastic mulch on May 15. Plots had two rows of 9 plants each spaced 60 cm apart. Of 1,152 plants evaluated, 97.7% were discarded due to early flowering or poor appearance and 27 outstanding late-flowering plants (2.3%) harvested on Sept. 12, cleaned and stored at 4°C.

Days to Flower Variability, 2001

To test the variability for days to flower (DTF) and selection effectiveness for late-flowering, arugula seed from the Italian source Detassis (total surviving population of 1664) and the LFS’s from 2000 (total surviving population of 178) were planted in 128-cell trays in the greenhouses of the Department of Horticulture and Landscape Architecture on Jan. 10, 2001. Plants in the trays were subjected daily to 16 hr of incandescent light (110 to 140 mmol m-2 s-1) immediately after germination for the next 3 moths. Plastic flats were placed underneath the plug trays and fertigated as needed. The number of plants flowering (at least one flower open) was recorded daily. After 65 days the remaining nonflowering plants were transplanted to 10 cm pots but flowering plants were continually removed. The experiment was terminated on Apr. 11, 91 days after planting.

At the end of the experiment, nine late-flowering greenhouse selections from the “cultivated arugula” (Detassis) and one from the 2001 LFS survived. Self-pollinations (open flower and bud) and intercrosses were made among the selections.

Field Evaluation of Late-flowering Selections, 2001

Seventeen LFS selections and three checks (‘Astro’, “arugula,” and the “cultivated arugula” from Florsilva) were evaluated in a grower’s field in Rockville, Indiana in 2001. The experiment was planted in a randomized complete block design with three replications on April 22, 2001. Plots were 0.9 m by 1.2 m and had 3 rows separated 15 cm from each other. Seed was planted with a hand vegetable seeder using a “radish” plate and the appropriate speed to drop 4–6 seeds per inch. Plots were grown organically and weeds controlled by hand. Water was supplied through a drip irrigation system. Plots were harvested on May 19, cutting the plants about 2 cm above ground. The plants were allowed to regrow and then harvested again on June 8, 2001.

USDA Germplasm Collection

To identify sources for late flowering and other horticultural characteristics, the Eruca sativa germplasm collection (163 accessions) from the USDA along with three entries (two of cultivated arugula and one of wild arugula) from Italian seed sources (Detassis, and Florsilva) were planted in the greenhouses on June 7, 2001 (Fig. 2). There were 128 seedlings of most accessions in the USDA collection and 256 of the commercial sources. The young plants were exposed to 16 hr of incandescent light for two months.

Fig. 2. Arugula plants of the USDA germplasm collection growing in the greenhouse. Notice differences in leaf size and shape.


Late-flowering Selection

A commercial source of cultivated arugula from Detassis showed large variability for days to first flower (Fig. 3), leaf patterns (Fig. 4) and stem color, and branching habit (Fig. 5). This variability is expected to permit progress using mass selection in the field. However, low natural seed set in the greenhouse makes greenhouse selections difficult to handle without a source of pollinating insects. We detected differences in pungency of selections but this could not be extensively investigated because of the use of pesticides for insect control in the greenhouse.

Fig. 3. Comparison of flowering distribution between the Italian source Detassis and late-flowering selections (LFS) of arugula.
Fig. 4. Leaf morphology variation in arugula.
Fig. 5. Stem color (green to purple) and branching variation in arugula.

Although our tests are preliminary we have demonstrated that selection for late blooming can be achieved using mass selection. For example, after a single cycle of selection for late flowering, only 9% of seedlings derived from late-flowering selections flowered within two months from planting, as compared to 35% of the original cultivar. Furthermore, in greenhouse evaluations, original selections proved to be 15% later than a comparable Italian cultivar (see below).

Comparison of Late-flowering Selections with an Italian Source

In the greenhouse experiment planted on Jan. 10, 2001, flowering in cultivated arugula from Detassis was compared with seedlings from late-flowering selections (Fig. 3). In arugula from Detassis, the first plants flowered 31 days after planting. Of a total evaluated population of 1664 plants, half flowered by 59 DAP. After 91 days, when the experiment was terminated, 14 non-flowering plants (0.8%) remained. The standard deviation of DTF for flowering plants was 12.3 days. In contrast, plants derived from late-flowering selections (LFS) started flowering at 47 days, 16 days later than the plants from Detassis. Of a total population of 178 plants, half flowered at 68 DAP. After 91 days, 6 plants (3.3%) remained. The standard deviation of DTF for flowering plants was 10.7 days.

Genetic gain cannot be calculated because the initial source population was not directly tested but our data indicates a mean genetic difference of 9 days for lateness of the selected population with the Italian source. Assuming no difference in the two populations (Johnny’s and Italian), this would indicate a 15% gain after one cycle.

Nine non-flowering plants from the Italian source and one from the LFS survived transplanting to 8-L pots. At flowering in June, intercrosses and self-pollinations were made. Intercrosses produced seed but only bud pollinations produced seed from self-pollination, indicating that self-incompatability is present in arugula and that it could be overcome by bud pollination. Selfings and hybrids are presently under evaluation.

USDA Germplasm Collection

Of the 163 lines received, five were excluded because of poor germination. The 158 lines remaining were from 15 different countries, but mostly from Pakistan (Table 1). Most entries had a rather low mean and small range for DTF. Pakistan, with the largest number of lines, had a low mean and range suggesting that this population is quite uniform for this trait. India, represented by 15 lines, had the highest mean (days to 50% flowering) and the widest range and appeared to be the best source of late flowering in the USDA germplasm. However, the two Italian commercial sources of cultivated arugula used as a control were by far the best performers for late flowering. They had a combined mean of 103 DTF, which is almost twice the 55 days mean of the USDA collection. It appears that for centuries Italian growers have been effectively selecting for late flowering. Despite this selection, there appears to be considerable variability in Italian arugula populations to make further progress.

Table 1. Mean and distribution of days to 50% flowering on arugula germplasm, grown under 16 hr photoperiod.

Origin No. of lines Days to 50%
USDA Collection of Eruca sativa
Afghanistan 3 47 32-73
China 1 58 39-76
Cyprus 1 51 35-64
Czechoslovakia 1 51 35-64
Egypt 2 52 37-64
England 2 43 32-73
Germany 2 57 37-76
India 13 64 28-150
Iran 9 47 28-73
Italy 1 59 42-76
Libya 2 57 42-76
Pakistan 114 48 28-115
Poland 1 71 46-87
Spain 1 59 42-80
Turkey 5 58 32-101
Mean and range   55 28-150
Commercial sources
Cultivated arugula      
Detassis, Italy 1 109 53-156
Florsilva, Italy 1 98 53-150
Wild arugula      
Florsilva, Italy 1 63 42-87

Field Evaluation of Late-flowering Selections

The field trial of 17 LFS and 3 checks was harvested at a very young stage on May 19 and the regrowth on June 8. However, further harvests were non-possible because the entire trial flowered (Fig 6). There was a significant difference in yield between entries for first harvest, second harvest, and total harvest but the high coefficient of variation (CV) made mean separation difficult (Table 2). Thus, only the top and lowest entry were significantly different from each other for total yield. The mean yield of the experiment (753 g per plot) is equivalent to a yield of 7.3 tonnes/ha.

Table 2. First, second and total foliage yield and DTF of late-flowering selections of arugula grown at Rockville, Indiana.

Line Harvest (g/1.08 m2)
First Second Total DTFw
7-1 407 ax 561 ab 968 a 57 a
10-1 316 ab 636 a 952 ab 55 a
7-4 295 ab 652 a 948 ab 60 a
10-18 292 ab 614 a 906 ab 57 a
10-4 389 ab 490 ab 879 ab 60 a
Arugulaz 264 abc 601 a 866 ab 61 a
6-1 311 ab 467 ab 778 ab 56 a
10-2 296 ab 481 ab 777 ab 59 a
1-2 307 ab 464 ab 771 ab 59 a
10-51 246 abc 509 ab 755 ab 59 a
8-1 210 bc 514 ab 725 ab 59 a
7-2 355 ab 360 ab 714 ab 59 a
7-3 273 abc 432 ab 705 ab 59 a
5-2 281 abc 398 ab 680 ab 57 a
10-21 230 abc 420 ab 650 ab 57 a
10-15 309 ab 322 ab 631 ab 57 a
1-1 269 abc 361 ab 630 ab 59 a
Italiany 272 abc 352 ab 624 ab 55 a
Astroz 100 c 520 ab 620 ab 55 a
10-6 244 abc 246 b 490 b 59 a
Mean 283 470 753 58
CV 34 36 31 6.3

zJohnny’s Selected Seeds, USA
yFlorsilva Seed Company, Italy
xMean separation by Duncan’s multiple range test
wDTF = Days to flower

Flowering in arugula seems to be influenced by cold temperature (vernalization), long days, and high temperatures. The vernalization effect is based on the observation that volunteer overwintering seedlings, originated from seed left in the field, flowered as tiny plants with only a few leaves in early spring. In this trial, flowering occurred from June 15 to 28 with a mean DTF of 58 days. However, there was no significant difference between lines. The lack of difference between lines may reflect a vernalization effect, because the experiment was planted on April 22, when night temperatures still dropped below 0°C. Cold temperatures may have induced flowering in all the arugula lines and checks, eliminating any genetic differences in photoperiod response. Flowering may have been delayed by the cutting of the plants at the first and second harvests on May 19 and June 8, respectively.


Arugula is a promising green for the Midwest with a spicy flavor that can compete with lettuce. Unlike lettuce, arugula is not bitter.

Arugula can be harvested from sequential plantings and it is possible to perform several harvests. Sequential harvest seems to delay flowering. We recommend planting at two to three-week intervals.

Early bolting can be reduced by selection for late flowering. Preliminary results indicate that mass selection for late flowering is effective.

The evaluation of variability of flowering response to 16 hr photoperiod of 158 accessions of the USDA germplasm collection and three Italian cultivars (two cultivated and one “wild”) indicated that the two Italian cultivated cultivars were twice as late as any of the accessions in the collection.