Parsnip webworm

“Parsnip moth” redirects here. Not to be confused with Agonopterix heracliana, a leaf-rolling moth that has carried this one’s old name since 1966.

The parsnip webworm or parsnip moth (Depressaria radiella) is a small moth of the family Depressariidae. Its caterpillars live under silk on the flowering umbels of parsnips and hogweeds, eat the flowers and the ripening seed, and then chew their way into the hollow stem to pupate. They eat nothing else. In its native and its introduced ranges alike, the insect is confined to the reproductive parts of two genera, Pastinaca and Heracleum;¹only the means of reproduction the second includes the giant hogweeds.

The moth is native to Europe. It reached North America by 1869 and New Zealand by 2004, both times by accident.²³ Its caterpillars thrive on furanocoumarins, the phototoxic defence of their food plants, which poison other insects, and its relations with wild parsnip have been studied as a case of coevolution since the 1980s. In North America, where the parsnip had grown without its enemy since the seventeenth century, the plant became more toxic once the moth caught up with it.⁴ The moth has been proposed for the biological control of Sosnowsky’s hogweed.⁵ The holes it leaves in hogweed stems shelter woodlice, earwigs and spiders for months after the moths have gone.⁶

Taxonomy

Names

In the last sixty years the moth has had three scientific names. Linnaeus’s Phalaena heracliana of 1758 mixed two species: he cited figures of a caterpillar that rolls leaves, now Agonopterix heracliana, and of one that feeds in umbels, this moth. For more than two hundred years the umbel feeder was known as Depressaria heracliana.⁷ In 1966 Bradley, who saw only two specimens in Linnaeus’s collection, both of them the leaf-roller, designated one of them as the type of the name. Heracliana passed to Agonopterix, and the parsnip moth became D. pastinacella, a name published by Duponchel in 1838, which “came slowly into use”.⁷ Karsholt and his co-authors later called the designation “a most inappropriate action”: it moved a well-known name from one species to another, and left the genus Depressaria with a misidentified type species.⁷the name changed owners

In a paper dated 2005 but published on 17 February 2006, they showed that the oldest available name was Phalaena radiella, given in 1783 by Goeze to a moth that Geoffroy had described from around Paris without a binomial name. A rival name published by Retzius in the same year was unavailable, because Retzius had not used binomial names consistently.⁷ They also fixed the moth as the type species of Depressaria, the genus that gives the family its name.⁷ American research kept to pastinacella until at least 2020,⁸ and the German lepidopterists’ wiki describes the episode as nomenklatorisches Chaos.⁹

Riley called the insect the “parsnip web-worm” in 1888, and in North America it is the parsnip webworm;¹⁰ in New Zealand it is the parsnip moth, and in German the Pastinakmotte.³¹¹

A native, briefly

In August 1869 the editor of The Canadian Entomologist, Charles Bethune, described the insect that had destroyed his parsnip seed as a new species, Depressaria ontariella. He judged the “ravager” to be “in all probability a native”, and named it after the province.² In the next issue he printed a letter from James Angus of West Farms, New York, who had reared the caterpillars “for many years” and took them for the British heracliana. Bethune conceded that it was “not at all unlikely” that the insect was imported, “like a great many more of the farmers’ and gardeners’ worst pests”, and offered to withdraw his name if it proved British: synonyms, he wrote, were becoming “such a nuisance”.¹² In 1873 Zeller found two females labelled ontariella to agree exactly with large European specimens; as Riley reported it, the species had “emigrated to America” and, after “a happy winter passage”, had found food for its young without difficulty.¹⁰ Bethune’s type specimens are probably lost, perhaps eaten by Dermestes beetles.¹³

Classification

Older American work placed the moth in the family Oecophoridae, and some papers of the 2000s in Elachistidae. A revision of the superfamily Gelechioidea in 2014 redefined Depressariidae as a family of its own.¹⁴

Description

The moth has a wingspan of about 19–29 mm; the sources differ slightly.¹⁰¹³¹⁵ It rests with its wings laid flat over its back, which gives it a flattened look;¹⁶ the German name of the family is Flachleibmotten, “flat-bodied moths”.¹⁷ The grown caterpillar is greenish yellow, with a shiny black head and black warts that each bear a bristle, and is 12–18 mm long.¹⁶¹⁰

All the European and North American specimens whose DNA barcodes were read in full shared a single sequence, except one from 2,600 m on Mount Terskol in the Caucasus. It differed at several bases and had a wingspan of 32 mm, and may belong to another species.¹⁵

Life cycle

The moth has one generation a year.¹⁵ The adults spend the winter under bark, in farm buildings or indoors.¹⁶¹⁸ Bethune saw them flitting about rooms and coming out from behind curtains in midwinter, usually mistaken for clothes moths,² and E. B. Southwick found one on a lace curtain on New Year’s night.¹⁸ Newly emerged moths do not breed until they have been chilled; in the laboratory they are given an “artificial winter” of about fourteen weeks at 10 °C.¹⁹²⁰ Southwick thought it “poor economy” for the moths to go into hibernation so early in the season.¹⁸

The eggs are laid singly, mostly on the undersides of leaves.²¹²² The caterpillars move to the flower buds and feed on buds, flowers and developing fruit inside a web spun over part of the umbel.²¹ Bethune described infested umbels shrinking into “shapeless masses of web and excrement”.² The caterpillars find the flowers by smell: octyl acetate, a scent of the flowers, attracts them, and octyl butyrate repels them.²³ There are six larval stages. In the laboratory the egg hatches in about four days, the caterpillar feeds for about 21 and the pupa takes about 13, some 38 days from egg to moth.²²

Webs

A caterpillar spends about a third of its time spinning, and puts almost 18 per cent of the nitrogen it eats into silk.²⁴a third of the day on fences The webs are defended. Of 63 staged encounters, 52 were started by the resident, and all the residents but one kept their webs; caterpillars turned out of their own webs moved into empty ones 87.5 per cent of the time rather than spin new ones.²⁴ The web seems to keep the peace. Caterpillars whose webs were taken away every day fought until they had re-established their “territories”, spun the most silk and made the lightest pupae; caterpillars kept alone spun the least and grew heaviest.²⁵ Early arrivals take over sections of an umbel, and most umbels hold a single caterpillar.²⁶ In captivity the caterpillars eat each other.¹⁰

In the stem

When fully grown, the caterpillars crawl down the plant, chew a hole about 7 mm across into the hollow stem, and pupate inside under a loose fold of silk.⁶²⁷ Bethune watched them eat the soft white lining of the stem “out of sight of all their enemies”.² Twenty caterpillars or more may feed on one plant;⁶²⁷ Dutch stems held from 1 to 21 pupae, 4 in the median stem,²⁸ and in Ontario in 1869 one “could hardly cut a stem” without cutting a caterpillar or a pupa.² The moths come out in mid to late summer: from 1 August in Ontario in 1869, and near Moscow at the end of August and in early September.²⁵

Food plants

The webworm feeds only on parsnips and hogweeds, and only on their flowers and fruit. The scientists who study it call it a “super-specialist”.¹¹⁵ Records from other plants of the carrot family are few. In America wild carrot was listed as a host, but Southwick found no caterpillars on it where it was abundant, and caged moths would not lay eggs on it.¹⁸ In Europe the main hosts are wild parsnip, common hogweed and giant hogweed,²¹ and near Moscow the moth also feeds on Sosnowsky’s hogweed.⁵

Its favourite in Europe is common hogweed. In a Dutch survey in 1997 webworms occupied 144 of 160 common hogweeds, 32 of 67 parsnips and 10 of 35 giant hogweeds, with means of 12.8, 3.0 and 1.6 caterpillars a plant, although the giant was the largest of the three.²¹ In the Netherlands parsnip is “rarely used”, since common hogweed is “greatly preferred”,²⁸ and the close partnership of parsnip and webworm that is familiar in North America is “exceptional in Europe”.¹ Moths raised on giant hogweed were about a tenth smaller than moths raised on common hogweed.²⁸ The giant’s seed held about three times as much furanocoumarin as the seed of common hogweed, and nearly three hundred times as much angelicin.²¹

Across much of North America the webworm’s only host is wild parsnip. In some 160 years there it has taken up a single native plant, the cow parsnip, Heracleum maximum.²¹¹⁵ The choice looks paradoxical: caterpillars survive less well on cow parsnip, and over four years the pupae in its stems were heavily taken, apparently by birds, while pupae in the tougher stems of parsnip escaped. The authors suggested that parsnip stems offer the moth an “enemy-free space”.²⁹ In the Rocky Mountains the moth and its parasite live on cow parsnip as high as 2,700 m.⁶

Chemistry and coevolution

Detoxification

The furanocoumarins of parsnips and hogweeds poison most insects. Xanthotoxin, which is toxic to caterpillars that feed on many kinds of plant, has no effect on the growth or survival of the webworm, whose gut enzymes, cytochrome P450s, are highly active against it and are switched on by it.³⁰ A caterpillar breaks down about 95 per cent of the xanthotoxin it eats, and passes some of it out through its silk glands.³¹ Webworms from eastern North America break xanthotoxin down 10 to 300 times as fast as other Lepidoptera.²¹

Each of the moth’s known enzymes handles only a few of the plant’s compounds. CYP6AB3 breaks down imperatorin, and a variant of it also myristicin, while CYP6AE89 deals with bergapten and xanthotoxin.³²³³⁸ A caterpillar cannot adjust the mix to its diet: its detoxification profile appears to be fixed in the individual, and varies from one individual to another.³⁴ When the protein in its food is cut, a caterpillar grows less but keeps its detoxification at full strength, “at the expense of growth”. On a diet with no protein at all, which cut growth by almost 80 per cent, xanthotoxin still switched on its enzymes almost threefold.³⁵defence is the last thing cut

Arms race

Wild parsnip resists the webworm with the same compounds. Its resistance is inherited: in an experimental garden four heritable furanocoumarin traits explained about three-quarters of the variation in resistance among families. Resistant plants, however, set less seed, and would be at a “competitive disadvantage” without the moth. The authors concluded that plant and insect had reached an evolutionary “stalemate”.³⁶ Protected from the webworm with insecticide, wild parsnips produced two and a half times as much seed.³⁷ Parsnips also set fruits without seeds. The caterpillars prefer them, although they grow more slowly on them, and the empty fruits may act as decoys that draw the attack away from the fruits that matter.³⁸empty shells attract the most interest

In 1998 May Berenbaum and Arthur Zangerl compared four populations in Minnesota and Illinois. The chemical types of plant and of caterpillar occurred in matching proportions, to an “extraordinary degree” in three of the four. They described the two as caught in a cycle of “escape” and “chase” between an “exploiter” and a “victim”, a cycle that may differ from place to place and turn “endlessly”.²⁰ Of twenty populations in Illinois and Wisconsin examined later, twelve matched; cow parsnip growing nearby made a mismatch much more likely.³⁹

Re-armament in North America

Parsnips were grown in Virginia by 1609 and were common there by 1630. In 2005 Zangerl and Berenbaum used herbarium specimens covering 152 years to follow what happened when the moth caught up with them. None of the 37 North American plants collected between 1850 and 1889 bore webworm damage, and their seeds held significantly less of four of the five main furanocoumarins than European seeds of the same period. Damage appears in the specimens after 1889. Between the periods 1850–1889 and 1890–1909 all five compounds rose significantly, and they went on rising. European seeds collected between 1819 and 2000 showed no such trend.⁴

1820–1889  ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒     31
1850–1889  ██████████▌          21
           · · · · · · · · · · · 
1890–1909  ████████████████     32
1910–1929  ██████████████▌      29
1930–1949  ███████████████      30
1950–1969  ████████████████     32
1970–2002  ███████████████████  38

The plants without their enemy had been saving. By a published model of seed yield, the difference in toxicity was worth about 0.127 g of seed per plant, some 23 per cent of the seed of a small plant and 3 per cent of that of a large one. Without a specialist, the authors inferred, the “high cost” of defence had selected for less of it.⁴ Invasive weeds are thought to spread by the “reallocation of resources from chemical defense into growth and reproduction”;⁴the peace dividend when the old enemy came back, the parsnip re-armed. The authors concluded that introducing a specialist herbivore “can increase noxiousness” of a weed, a point for anyone planning the biological control of a poisonous one.⁴

New Zealand

Parsnip had been recorded in New Zealand since 1867, and common hogweed since 1939, when the moth was identified there on 2 February 2004.⁴⁰³ It was already too widespread around Port Chalmers and Dunedin to be eradicated. The Ministry of Agriculture and Forestry classed it as a “Case 3” incursion, one established “well beyond” the point at which action is cost-effective.³⁴⁰written off By February 2006 the moth had spread north to Christchurch.⁴⁰

The reunion was studied as it happened. In 2006 the webworms cost about half the plants in two populations half or more of their fitness, and in 2007 three-quarters of the plants in two populations set no seed at all. Selection on the plants’ chemistry changed, favouring what the authors called “trait remixing”. Uninfested New Zealand parsnips produced more octyl acetate, the scent the caterpillars follow; the authors suggested that plants freed from their specialists may evolve more of the scents that attract them, as well as fewer defences.⁴⁰ After three to six years of attack, the infested populations had evolved larger plants, better able to bear the damage, while their chemistry had not changed.⁴¹ Long-term control of a weed by a coevolved specialist, the authors of the first study concluded, “may be an elusive goal”.⁴⁰

Natural enemies

In Europe the webworm’s most important enemy is probably a tiny wasp, Copidosoma sosares.⁶ The female lays in the moth’s egg, and her own egg divides as the caterpillar grows, producing from 10 to 300 genetically identical wasps in a single host; mean broods in a Dutch survey were between about 136 and 237.²¹ The parasitised caterpillar becomes a “mummy” in the stem, and the wasps leave by the webworm’s own hole.²⁸ In one Dutch season 28,571 wasps emerged from the mummies collected.²⁸ In Europe its attack rates “often exceed 80%”.²¹

The plant’s poison reaches the wasp. On a diet rich in xanthotoxin, broods were more than a fifth smaller. The xanthotoxin that the caterpillar fails to break down reaches its blood, where the wasp embryos develop, at four times the usual concentration, and neither the embryos nor the wasp’s precocious larvae can break it down.⁴²

In eastern and midwestern North America the webworm has been “largely free” of important enemies. In two seasons in Iowa, parasites took 2.7 and 0.8 per cent of the caterpillars, and fifteen years of sampling in Illinois found no higher rates.²¹²² Copidosoma sosares has since been found in eight western states, indistinguishable from Dutch wasps. There is no record that anyone introduced it, and it probably came with its host.⁴³ It now appears to be establishing itself in Wisconsin and Illinois.²⁸

Other enemies have been recorded over the years. The potter wasp Eumenes fraterna prises caterpillars out of their webs and stocks its cells with them;¹⁸ a hairy woodpecker visited Bethune’s parsnip stalks every day to peck out caterpillars and pupae;² and in England in 1912 the common earwig destroyed “hundreds of pupae”.⁴⁴ In 1891 a disease killed about five-eighths of the caterpillars along one ditch, and at a meeting of economic entomologists the next year Forbes suggested that “much might be accomplished” by spreading it.¹⁸

Holes

In the Netherlands, hogweed stems without webworm holes were “always free” of other arthropods. Stems with holes often held many woodlice, earwigs and spiders for three or four months, until the plants died.¹¹ In a survey of 91 perforated plants in the Netherlands and the United States, more than twenty woodlice lived in each of half the stems examined at Wageningen, earwigs in more than four-fifths of the Dutch stems, and up to sixty earwigs in a single stem at Santa Fe, New Mexico; stems there still held earwigs a year later.⁶ The authors called the moth a habitat “facilitator”, and in America a “re-facilitator”; a later study called it an “ecosystem engineer”.⁶²⁷the builders move out, the tenants move in

In 2021 researchers near Leiden punched holes about 7 mm across into the stems of 200 hogweeds. Woodlice and earwigs moved into them as they do into the moth’s. Earwigs eat orchard pests, and the authors proposed such holes as a way of attracting them.²⁷

Distribution and spread

The moth is native across Europe, from the British Isles and Scandinavia to the Caucasus,¹⁰⁴⁵ and in Russia is recorded from the European part, Siberia and the Far East.⁵

Its first North American record is Bethune’s, from Ontario in 1869, although a correspondent in New York said the same year that he had known it “for many years”.²¹² Charles Valentine Riley found it “extremely common” in wild parsnip at Kittery Point, Maine, in 1883, and by 1888 considered it “now well seated in the Eastern United States”.¹⁰ Then it spread west. The earliest records in collections include Illinois in 1900, Utah in 1907, Portland, Oregon, in 1914, and Arizona and British Columbia in 1925; by 1932 it was reported from every province of Canada.¹⁵¹³ These are first records, not dates of arrival.

A pest of seed crops

For growers of parsnip seed the moth was a disaster. Bethune lost the seed of a garden bed he had promised to his neighbours: “the prospect of seed was utterly and entirely gone”.² In June 1888 the seedsmen D. Landreth & Sons of Bristol, Pennsylvania, sent Riley infested seed stalks. They had tried kerosene emulsion, whale-oil soap, Paris green and other remedies, “all to no effect”.¹⁰ Southwick called the wild parsnip “a curse among the farmers”, and thought that if it were grown for seed the damage would cause alarm.¹⁸ In north Durham in 1912 many seeding parsnips were killed outright.⁴⁴

During the First World War vegetable seed was short, “consequent upon conditions in Europe”, and farmers in Nova Scotia planted parsnip seed plots of their own. Many complained of serious damage, and an attempt to grow parsnip seed on the college farm at Truro in 1914 failed. Cutting and burning the seed heads would have destroyed the crop, and the wild host plants were “too numerous” to destroy.¹⁶

Biological control

The moth was not introduced on purpose: it reached North America and New Zealand by accident, and New Zealand treated it as a pest.¹³ Its reunion with wild parsnip in North America nevertheless resembles a classical biological control release, and has lasted about twice as long as the oldest programme of its kind against a weed in the continental United States, the release of Chrysolina beetles against St John’s wort in 1944.⁴ Its results, a weed that became more poisonous in America and in New Zealand adapted within a few years, have been offered as a caution to anyone planning such a programme.¹⁴⁰

Between 2002 and 2005 a European project searched the Caucasus for enemies of giant hogweed. It found the webworm among the most abundant of the insects that feed on several plants of the carrot family, in the native and the invaded range alike,⁴⁵ and scientists of the project thought seed feeders such as Depressaria perhaps the most promising candidates.⁴⁶ But no enemy proved specific to giant hogweed: all attacked other hogweeds, and most attacked parsnip.⁴⁷ The webworm, in any case, was already in Europe. In July 2015, at Burscheid in the Rhineland, at least twenty of its caterpillars ran out of the hollow stems of two giant hogweeds that were being cut up and bagged for disposal. The authors of the report believed it to be the first record on the plant.¹⁷ It was not: webworms had been counted on giant hogweed in the Netherlands in 1997.²¹

Near Moscow, M. G. Krivosheina found webworm caterpillars eating the ovaries of Sosnowsky’s hogweed and pupating in its stems, and judged the moth and the noctuid Dasypolia templi the most promising agents against it. Umbellifers in the Moscow region, she argued, are grown mostly for their greens and roots, not for seed, which lowers the risk to crops.⁵ Moths raised on Sosnowsky’s hogweed emerged a week or so earlier than moths raised on parsnip.⁵

See also

• Furanocoumarin § Arms race
• Enemy release hypothesis
• Common hogweed
• Giant hogweed
• Control § Biological control