Index | Search | Home | Table of Contents

Witt, M.D. and H.D. Knudsen. 1993. Milkweed cultivation for floss production. p. 428-431. In: J. Janick and J.E. Simon (eds.), New crops. Wiley, New York.

Milkweed Cultivation for Floss Production

Merle D. Witt and Herbert D. Knudsen

    1. Study I--Line Ecotypes
    2. Study II--Hybrid Ecotypes
    1. Crop Establishment
    2. Pod Components
    3. Floss Production
    4. Flowering
  5. Fig. 1
  6. Fig. 2

Field plot studies of milkweed, Asclepias speciosa and A. syriaca, as a row crop, were initiated in 1985 at Garden City, Kansas. Potential cultivars and interspecies hybrids were evaluated for their floss production potential. Milkweed floss is usable as a hollow, cellulose, insulative, batting fiber and competes readily with the more traditionally used white goose down, which costs $66.00/kg. Fifteen lines of these two species yielded a 4-year annual average of 212 kg floss/ha from dried pods that averaged 21.8% floss. Thirty-five interspecies hybrids yielded a 4-year annual average of 350 kg/ha from dried pods that averaged 22.7% floss. Approximately 227 pods were required to produce a kilogram of floss. In view of the high-value application of floss and a projected market use of 1.27 million kg/year, Asclepias species appear to be worthy of further crop development.

The genus Asclepias is composed of approximately 140 species worldwide (Woodson 1954, 1962) with 108 species identified in North America, where this dicotyledonous plant is a native. Asclepias species are generally considered as being persistent, perennial, hardy weeds containing cardiac glycosides that are toxic when ingested by livestock (Muenscher 1975). Because of the negative aspects of milkweed as a pest, research efforts have most often dealt with techniques associated with control of these resilient plants (Timmons 1946; Wyrill and Burnside 1976). At the same time, milkweed has been of interest to agriculturalists for many years because of its potential economic value as a new crop (Stevens 1945; Moore 1946; Berkman 1949).

Various plant components have been identified as having potential uses, including insulative and absorptive materials made from the floss (Adams et al. 1984), and liquid fuel and rubber (Buchanon et al. 1978), as well as oil and polymeric hydrocarbons (Adams et al. 1987), from the latex sap. The only plant component used in quantity to date is floss, which emerged as the most feasible alternative to replace kapok when imported supplies during World War II were halted. The suggestion of such a potential use (Groh and Dore 1945) was followed quickly by a report of floss being collected and used as a kapok substitute (Gladfelter 1946). Eventually 11 million kg of pods were consumed to fill 1.2 million "Mae West" life jackets.

The currently preferred batting product of high insulative quality is goose down, but it costs $66.00/kg to import from China. At present, milkweed floss is being utilized to blend with white goose down or synthesized polypropylene in insulative batting for such products as comforters, sleeping bags, and arctic apparel.

Inadequate information exists on the production potential of milkweed as a cultivated crop. Initial interest suggested that if milkweed is to be considered as a crop plant, attention should be given first to selection of superior strains (Stevens 1945). Naturally occurring interspecies hybrids with variable characteristics were soon reported (Moore 1946; Nicholsen and Russel 1955). However, we found no evidence that replicated production evaluations of milkweed germplasm strains were ever conducted. This research effort was designed to evaluate the agronomic potential and variability of 50 collected milkweed ecotype strains and hybrids of Asclepias speciosa (showy milkweed) and Asclepias syriaca (common milkweed) when grown in monoculture as a new fiber-producing row crop. By providing an alternative cash crop, milkweed farming could strengthen the agricultural diversity and economy of the western Great Plains.


Two field studies were established in a small, leveled flood irrigation basin and treated similarly. Plots were established from seed collected through efforts funded previously by the Standard Oil Company of Cleveland, Ohio. Seeds that had been repeatedly water soaked and dried to break dormancy were planted in 76.2 cm rows in a Keith Silt Loam type soil on May 30, 1985. A John Deere 71 flex-planter with cone seeding units and disk openers with depth bands was used to plant all plots with seed placed 1.27 cm deep. A seeding rate of 5.6 kg/ha placed approximately 1,111,500 seeds/ha into the seedbed. Stands were thinned to 10 cm between plants, when they had reached 8 cm in height on July 10, 1985. Two adjacent field studies were established.

Study I--Line Ecotypes

Fifteen randomly collected entries from 10 states (Colorado, Idaho, Kansas, Maine, Montana, New Mexico, North Dakota, Oregon, Utah, and Washington) included 12 showy milkweeds, two common milkweeds, and a single interspecies hybrid from two parent entries. The accessions were seeded in a randomized complete block design using three replications with three row plots and 3.04 m row length.

Study II--Hybrid Ecotypes

Thirty-five entries originating in Kansas and eastern Colorado and identified taxonomically by Kansas State University herbarium personnel as naturally occurring showy milkweed x common milkweed crosses were included. Because of limited seed quantity, these were seeded in a randomized complete block design using two replications with single-row 3.04-m-long plots.

All plots were overhead irrigated for 3 weeks after sowing to aid seedling emergence. Thereafter, irrigation was applied to plots 2 or 3 times each summer. Ammonium nitrate fertilizer granules were broadcast each season at 67.2 kg N/ha. Milkweed beetle (Tetraopes tetraphthalmus Forst.) control was initiated in 1989, using Malathion 25% WP at 0.56 kg/ha. Trifuralin applications were incorporated between rows each year in early May for weed control.

Summer harvest of pods began each season when approximately 10% of the pods had burst open. Ten representative pods from each entry were separated into floss, seed, and empty shell and weighed. The resulting dry weight proportions were used to partition total harvested pod weights per plot into calculated components from 4.18 m2 harvest areas for the line ecotypes and from 1.39 m2 harvest areas for the hybrid ecotypes. Harvest dates over the years ranged from July 24 to Aug. 15.


Crop Establishment

Establishment was excellent, with seed germination as high as 80% resulting in uniform stands. Initial year stands were totally vegetative and reached average heights of 20 to 48 cm by the first killing freeze. Starting in 1986 and thereafter, the crop began to grow each spring from underground buds in late April, flowers appeared in early June, and pods were harvested 6 to 8 weeks later.

Plants were vigorous and developed extensive root systems with numerous viable root buds. Roots spread outward as much as 2 m from the unbordered perimeter rows by the end of 1986. Root spreading within and between plots was much more restricted, apparently due to plant competition and between-row cultivation that excised unwanted shoots so that the 76.2 cm row spacing could be maintained.

Pod Components

Harvested plant pods had a similar ratio of component parts for both species and for the interspecies hybrids (Fig. 1). About 24% of the dried weight of pods was floss. The seed comprised the largest portion of the pod mass (about 40%). The shell component of 36% included the outer wall (34%) and the central placenta (2%). The exterior shell surface generally displayed more roughness with the showy milkweed species than did the shell of the common milkweed pods. However, shell wall thickness was similar between species.

Floss Production

Average annual productions of floss by the hybrid ecotypes and by the line ecotypes during the 4 years are indicated in Fig. 2. The hybrid ecotypes as a group averaged 349 kg floss/ha annually over the 4 years. The hybrid ecotypes consistently outproduced the line ecotypes. A direct statistical comparison between the two studies cannot be made. However, in comparison to line ecotypes, the hybrid ecotypes bloomed an average of 4 days later, stood 5 cm taller, and produced 58% more pods per area with 71% more pods per stem and 65% more floss per unit area.

Of the line ecotypes, the 12 showy milkweed ecotypes averaged 207 kg floss/ha, and the two common milkweed ecotypes yielded a similar level of 187 kg floss/ha over the 4-year period. Floss yields of the line ecotypes varied with differing state origins but did not appear to be highest from any particular state or region.

Yields of harvested floss were high in 1986 and 1988. Yields were hampered in 1987 and 1989. On June 16, 1987, during flowering, hail destroyed approximately 80% of the potential fruiting sites. In 1989, stands were reduced approximately 50% due to larvae of the milkweed beetle which had fed and tunnelled into the plant's root system following the 1988 harvest. Diseases were not a serious problem, although black bacterial leaf spot, leaf rust, and pod blight infections were observed late in the season in 1985 and 1986.


Milkweed is an obligate outbreeder, so that cross pollination is essential. Isolated colonies of wild plants were observed to bear few fruits in comparison with plantings of several colonies close together (Stevens 1945). We had hoped to provide adequate cross pollination in this study by rearing 50 different milkweed acquisitions within a 600 square meter area. The close proximity of these variable strains should have provided adequate pollen for maximum pod production.

Each single plant stem usually had about four flower clusters with about 40 flowers/cluster. Thus, there were about 160 flowers with two ovaries each, for a potential of about 320 pod sites/stem. However, this theoretical number was never approached. The best year (1988) allowed an overall average of 1.56 pods/stem for the variety ecotypes and an overall average of 2.3 pods/stem for the hybrid ecotypes.


As with all new crops, development of milkweed is a high risk, long-term venture. At the same time, the need for alternative crops to overproduced crops has never been greater. A market for a product is a crucial ingredient in the successful development of a new crop. Fortunately, the thermal impedance level of milkweed floss equals that of the well known goose down. This helps provide an already developed market and a major incentive for addressing remaining considerations in production and processing. Additionally, the production, harvesting, and floss extraction procedures and equipment needed by a new milkweed industry closely parallel the procedures and equipment already in use by the well established cotton industry. In this study, we have produced floss yields considerably beyond the 56 kg/ha minimum that we felt would be adequate to compete with goose down. The transition to greater usage of domestically grown milkweed floss in place of imported goose down would be a positive step, with substantial benefits for both consumers and producers.


Fig. 1. Components of a typical dried milkweed pod. Proportions did not differ significantly between A. speciosa, A. syriaca, and interspecies hybrids between them.

Fig. 2. Average floss yield of 35 hybrid ecotypes and 15 line ecotypes.
Last update September 15, 1997 aw