As urban green areas become better recognized for improving quality of life, with parks and lots replanted with native species, and urban/community gardens multiplying, more attention is being paid to species within cities [1]. Given their great diversity and ecological significance, as pollinators in particular, arthropods are a major focus besides the plants, or at least should be [2]. For example, in a 4-year survey of the community gardens in the Bronx and East Harlem in New York City, 54 species of bees were found, some 13% of the entire New York State fauna [3]. Among these 54 bee species, 19% were exotics.

Despite the loss of most native species [4], urban areas can surprisingly yield even new species. In 2002, for example, a new genus of tiny millipede was found in New York’s Central Park. The millipede was never officially described, although a name (Nannarup hoffmani—a nomen nudum) appeared in the New York Times article that announced the discovery [5]. As expected, many new urban species are actually introductions; “Nannarup”, for instance, putatively belonging to an Asian lineage, was probably introduced with soil from plantings. Newly discovered urbanites, however, can be indigenous, and familiar urbanites can be cryptic new species. Perhaps the most remarkable example of the latter is the recent report of a new morphocryptic species of leopard frog, Rana kauffeldi Feinberg et al., from northern New Jersey, southeastern (Hudson) New York, and one of the boroughs of New York City, Staten Island [6, 7]. Identification of this unique frog was based on nuclear and mitochondrial DNA sequences, mating calls, and some morphological features.

Here, we present two new, equally remarkable examples of hidden and unexpected biodiversity, but which involve drosophilid flies, thriving amidst one of the most populated urban centers in the world, Los Angeles, California. Drosophila is such an intensively studied eukaryote that it is perhaps the sole insect whose genus name is familiar to much of the general public. In an area housing over 1,000 people/km², it is astonishing that a survey would uncover two unusual drosophilid species, one previously collected only once (in Central America), the other known only from the Southern Hemisphere. Neither of these drosophilids was found by a Drosophila survey in Los Angeles four years earlier [8], in which another Drosophila new to California was documented, D. bifurca Patterson and Wheeler (albeit a species previously known from the southwest U.S. and northern Mexico) [9]. Notably, our new survey used Malaise traps, which rely on passive interception of flying insects. In contrast, other surveys (e.g., [8–10]) relied on banana-yeast baited traps, traditionally used to capture frugivorous drosophilids.

The common names for Drosophilidae, “fruit flies” or “pomace flies”, belie the true ecological diversity of the family. In fact, the range of larval hosts of drosophilids is extraordinary [11], and one of the more common lifestyles involves breeding in flowers [12]. An exemplar of flower-breeding drosophilids has been the Drosophila flavopilosa species group, one topic of our report, a lineage of 17 described species and at least 10 undescribed ones that occur throughout the neotropics from Veracruz, Mexico to central Chile, and the Caribbean islands [13–24]. The other discovery reported here is a highly unusual member of the Drosophila melanogaster species group. While originally thought to be found only in Eucalyptus flowers and therefore native to Australia and introduced into South Africa and Madagascar [25, 26], the presence of D. flavohirta in other myrtaceous genera native outside of Australia opens the question of the geographical origin of this species. As predicted [26], Drosophila flavohirta Malloch appears to be spreading to parts of the world where eucalypts have been introduced and cultivated; this report is the first known New World occurrence.

Our report of these two species has a primarily ecological and systematic emphasis. While it is highly unlikely that Drosophila gentica can be a pest on tomato, potato, peppers or other solanaceous crops, reporting their North American occurrence is important since some drosophilid introductions have had unexpected consequences. The most dramatic such consequence concerns the spotted-winged fruit fly, Drosophila suzukii (Matsumura). For many years this fly was a well-known major pest of cultivated fruit crops in its native Japan [27]. It appears to have been introduced into California in approximately 2008 and within three years spread throughout the southern half of North America up to Ontario [28], as well as into southern Europe [29]. It is a serious agricultural pest in its introduced range, particularly of small, soft berries and drupes [30], and it is even common in hardwood forests of the eastern U.S. (D. Grimaldi, pers. observ.). It is now also established in southern Brazil [31]. Given the temperate climate of Japan, establishment of D. suzukii in North America and Europe is not surprising, although the speed of its spread was remarkable.

Another important drosophilid invasion was by Zaprionus indianus Gupta, native to Africa, the Middle East, and India, where it is a highly polyphagous breeder in fermenting fruits. It was first detected in the New World in São Paulo, Brazil in 1999 [32] and had spread to Panama by 2003 and Florida by 2005 [33]. Now it is a significant agricultural pest in Brazil, particularly on figs (Ficus sp.), and it even appears to be adversely affecting populations there of native and other introduced frugivorous drosophilids [34]. Specimens are occasionally found as far north as New York City, although it is unknown if the species is breeding there since the species seems limited to tropical and subtropical climates.

One of the main features of Drosophila suzukii and Zaprionus indianus, indeed of all major invasive drosophilids—like Drosophila melanogaster Meigen, simulans Sturtevant, busckii Coquillett, repleta Woollaston, hydei Sturtevant, and funebris (Fabricius)—is their polyphagy and fecundity. Each of them can breed quickly on a wide array of fresh, damaged, or decaying fruits and organic substrates (e.g., [10]). As invasives, Drosophila gentica and D. flavohirta are unique in their host specialization and apparent low fecundity, laying a single large, very mature egg at a time, a common feature of anthophagous Drosophilidae. Drosophila gentica, in fact, was the second most abundant species in our Malaise trap samples from Los Angeles, after the common, cosmopolitan invasive Drosophila simulans. Drosophila flavohirta in South Africa actually appears to be a competitor with Apis mellifera Linn., the presence of fly larvae in the flowers making them unattractive to bees and negatively impacting eucalypt honey production [25, 35–37].

Specimens of both Drosophila species were originally collected as part of the BioSCAN project; D. gentica was subsequently found also on host plants. BioSCAN is an innovative project based at the Natural History Museum of Los Angeles County designed to survey urban biodiversity using a collecting method, Malaise traps, usually deployed in more natural habitats, while also engaging the general public in the sampling program. At each of 29 sites across the City of Los Angeles, California, a Malaise trap and weather station were set up (Fig 1). Almost all sites are backyards of private citizens who volunteered to host traps for the study; additionally, one is a school, one is the Natural History Museum of Los Angeles County (LACM) “Nature Garden”, and one is a community food garden. Express, written permission for the set up and maintenance of the traps by the landowners is on file at the Natural History Museum of Los Angeles County. No endangered or threatened species were involved. Geographic coordinates are given in the Material Examined section of the species treatments, summarized in the map in Fig 1. Malaise traps are tent-like structures made of a very fine-screened fabric, into which flying insects are intercepted and collected (Fig 1B). The traps can be set up for more than a year and emptied on a regular basis; they often collect insect species rarely captured using hand netting or other traditional methods. The diversity of focal groups of insects in the BioSCAN Malaise traps is being analyzed in their association with various indices of urbanization and land use to assess the effects of the city on the insect fauna. Additionally, besides surveying focal taxa, samples are screened for unusual taxa such as the Drosophila reported herein. Our new Drosophila records from the Malaise traps were collected and identified in the first three months of sampling of the BioSCAN project, during the winter when insect populations and diversity are lowest. Therefore, further unrecorded species are expected in summer samples.

All measured specimens were point mounted. In contrast to older specimens, new specimens were preserved in 70% ethanol treated with the solvent hexamethyldisilixane (HMDS), which preserves the color of specimens with more fidelity, also making them more fully distended and cleaner. Older museum specimens dried directly after capture or straight from ethanol are slightly greasy and appear slightly darker. Measurements of point-mounted specimens were generally made at 60-100X using a Nikon SMZ 1500 stereoscope with a Nikon DSRi1 digital camera and NIS Elements software; error range is approximately ± 0.01 mm (including the variation due to positioning the specimen). Standard measurements and ratios were made as given in [38]. A total of 43 measurements were made for each of five specimens from LA, five paratypes of D. gentica from El Salvador, and the four specimens of “gentica-like” from Jamaica (in [13]). All six specimens of D. flavohirta from LA were measured, and compared with measurements of five randomly sampled specimens collected 28 December, 2010 by S. McEvey in Stroud, New South Wales, Australia, off of Syzygium flowers (32.4079°S, 151.9672°E). Variation in measurements and ratios between samples of native and introduced specimens are given in the redescriptions only where significant (p<0.05) as based on a pairwise ANOVA for equal or unequal sample sizes. Male and female terminalia were dissected in representative specimens from all localities. Terminalia were macerated in warm 10% KOH, rinsed in water and 70% ethanol, dissected in glycerine using fine tungsten needles, and mounted on microscope slides in glycerine jelly for observation at 100-400X.

One of us (L. T.) collected flies on inflorescences in the Los Angeles area between 11–17 July, 2014, to assess potential hosts of the introduced species. The areas visited were: Mildred E. Mathis Botanical Garden (34°03'54.4"N, 118°26'28.4"W), the University of California, Los Angeles campus (34°04'12.2"N, 118°26'34.3"W), Tongva Park (34°00'41.4"N, 118°29'38.3"W), Palisades Park (34°00'00.8"N, 118°29'08.7"W to 34°01'32.8"N, 118°30'50.5"W), Los Angeles Arboretum and Botanical Garden (34°08'28.6"N, 118°03'10.4"W), Huntington Gardens (34°07'51.0"N, 118°06'52.0"W), and Rivas Canyon Park (34°03'25.0"N, 118°31'01.1"W). Flies were collected from Cestrum as well as various taxa of monocots, especially palms (Arecaceae), our thoughts at the time being that the species later identified as D. flavohirta might be a palm-breeding species. Although Cestrum sp. was found in two LA localities, only an anthomyiid and Thaumatomyia (Chloropidae) were found on these. Collecting by B.B. on Cestrum nocturnum in Los Angeles also found no D. gentica visitors. A species of Myrtaceae was sampled in LA (Melaleuca nesophila, or “showy honey myrtle”, native to Australia), on which only Milichiella sp. (Milichiidae) was found but no D. flavohirta. Thaumatomyia sp. was also found on Trachycarpus (Arecaceae) and Polianthes tuberosae (Asparagaceae); and another chloropid (Fiebrigella sp.) on Brahea nitida (Arecaceae). Anthomyiidae were collected from Yucca (Asparagaceae)(as was a specimen of Leucopis: Chamaemyiidae), and Dietes iridoides (Iridaceae). Another genus of Milichiidae (Milichia sp.) was found on Brahea brandegeei (Arecaeae).

One of us (M.T., assisted by W. M. Baca) searched specifically for D. gentica on Cestrum species at the University of California, Davis arboretum, the University of California Botanical Garden, Berkeley, and the San Francisco Botanical Garden (SFBG). On 30 July, 2014, D. gentica was observed in the SFBG (37°46’1.38”N, 122°28’11.45”W) by W. M. Baca on Cestrum aurantiacum. Several adult flies were collected and returned to U.C. Davis along with C. aurantiacum flowers. Adult D. gentica emerged from the collected flowers. Hence, the breeding range of D. gentica within California extends northward to at least San Francisco. Months of systematic sampling on flowers of Cestrum and various Myrtaceae throughout California are required to thoroughly assess host use and geographic distributions of the two drosophilid species.

In order to genetically determine the species identity of D. flavohirta in California, DNA was compared with specimens collected by S.M. in Stroud, Australia (locality data given above). DNA was extracted from a female flavohirta from sample #15778, and a female from Stroud, Australia, using standard ethanol extraction from which COI partial coding sequences were obtained and compared. Universal primers 5’-CCTACAGGAATTAAAATTTTTAGA-3’ and 5’- TCCAATGCACTAATCTGCCATAT-3’ were used to amplify a ca. 600-bp region of the COI CDS [39]. PCR conditions were as follows: denaturation at 94°C for 30 sec., annealing at 55°C for 30 sec., and elongation at 72°C for 30 sec for 35 cycles. Sequence alignments were performed using MUSCLE 3.8 [40].

Alignment of the LA and Stroud, Australia sequences showed 98.2% identity. This level of divergence within the amplified region is typical of intraspecific divergence within various species of the D. melanogaster species group. For example, a standard nucleotide BLAST of the D. simulans sequence, homologous to that amplified from the flavohirta samples, revealed nucleotide identities ranging from 98.0–100% amongst populations of simulans (the homologous region was obtained from D. simulans isolate RU259 complete mitochondrial genome, accession number AF200849.1). Additional collections and analyses are needed to comprehensively determine divergence of D. flavohirta populations and the placement of flavohirta within Sophophora [41–43], as discussed below with some new molecular data (S1 Text).

Repositories for the specimens are the Los Angeles County Museum, California Department of Food Agriculture (Sacramento, California), The Australian Museum (Sydney), and American Museum of Natural History (New York).

Flower-breeding Drosophilidae are very rarely collected in natural areas by sweeping, in Malaise traps, or by other general methods used for collecting insects; they are virtually never captured using fruit baits (D. Grimaldi, S. McEvey, pers. observations). For example, among the thousands of drosophilids amassed during the Zurqui All Diptera Biodiversity Inventory project in Costa Rica (B. Brown, Principal Investigator), collected using all general and many specialized field methods (including Malaise traps), there are only eight flavopilosa-group specimens. Flies in the flavopilosa group readily avoid detection even though they occur in abundance at flowers of Cestrum spp. during appropriate seasons. That substantial numbers of Drosophila gentica have been captured in Malaise traps from various spots in Los Angeles, and the species has been found in Los Osos and San Francisco, indicates that the populations are significant in size and well established. Several females from the Malaise samples contained mature eggs, confirming that these are breeding populations, which can be quite large on very local scales. Brncic [12], for example, calculated the density of adult Drosophila flavopilosa Frey in an area of Cestrum parqui scrub in central Chile, based on larval infestation rates of hundreds of flowers. Allowing for 90% pre-adult mortality, Brncic estimated the density of adults exceeded 30,000 flies per 100 m².

So far as known, all authoritative host records of the flavopilosa group indicate that species are restricted to flowers of Cestrum (Solanaceae) (references above), a Neotropical and warm temperate genus of some 175 species of trees and shrubs [52]. Santos and Vilela [24] reared two flavopilosa-group species from Sessea brasiliensis in São Paulo state, southern Brazil, the only host record from this genus of 18 species. Brncic [12] mentioned (pg. 361) that “in Central and tropical America, it seems that other members of the flavopilosa group may utilize other flowers,” for which he cited [53]. Unfortunately, several of the drosophilid identifications in [53] are known to be incorrect (e.g., the common, cosmopolitan species Drosophila immigrans Sturtevant is described as D. flexipilosa Pipkin, 1964); thus, identifications in that paper are suspect. Indeed, there are no flavopilosa group flies in Pipkin’s material in the U.S. National Museum of Natural History (D. Grimaldi pers. observ.). In a large survey of flower-breeding Drosophilidae from southern and eastern Brazil [54], flowers were collected from 125 plant species in 47 families. No Cestrum or Sessea were collected, though flowers of other Solanaceae were, including Brugmansia, Brunfelsia, Datura, Petunia, Solanum, and Streptosolen. In that study there were found 28 species of Drosophilidae breeding in 56 of the flower species and 18 of the families, but no specimens of the flavopilosa group. It is quite likely that these flies are indeed restricted to breeding in Cestrum and Sessea. The studies conducted and reviewed by [12] and by [24] indicate that some species of these flies breed in multiple species of Cestrum, others appear to be strictly monophagous. As noted, we found large numbers of D. gentica in the flowers of a Cestrum hybrid in Los Osos, CA (breeding them from the flowers was not attempted), and this fly was actually breeding in Cestrum aurantiacum in San Francisco. Cestrum aurantiacum is introduced into California, its native distribution being Mexico, Costa Rica, Guatemala, and Venezuela. More extensive field collections are needed to clarify the host and geographic range of D. gentica in California.

Host specialization of flavopilosa-group flies on Cestrum reflects the chemistry of this plant genus. The flowers, commonly called “sweet jessamine”, are highly fragrant, and the plants are notoriously toxic, particularly the reproductive structures. Ingestion of foliage, fruits, and flowers can be lethal to humans and grazing livestock [55–57]. Among the toxic types of compounds in Cestrum are glycosides and alkaloids, the former of which are implicated in acute hypercalcemia, calcinosis, and strynchine-like poisoning in ungulates [58–61]. Brncic [12] mentioned that “filtered homogenates” of Cestrum tissues kill the larvae of highly polyphagous, common species like Drosophila melanogaster, simulans, immigrans, hydei and funebris.

Cestrum parqui L’Herr, the most ubiquitous species of the genus in California, was introduced from southern South America into California, and is reported to occur in Los Angeles and Santa Barbara counties, as well as several counties around and including San Francisco [62]. In the Jepson Herbarium at Univ. California, Berkeley, there are specimens collected in 1893 from Napa City (Napa Co.), 1933 in Amador Co., and 1961 in Santa Barbara Co. Univ. California [63]. Interestingly, there is no mention of C. parqui in the California state flora [64], nor in the first edition of the Jepson Manual, although the second edition [65] indicates it is a “waif/garden escape” species. In Los Angeles Cestrum parqui is not widely planted but is a weedy species that readily disperses (P. Rundel, UCLA Bot. Garden, pers. comm. to LT May 2014). The genus Cestrum is not mentioned at all in [66]. Cestrum parqui has also been introduced toTexas [67].

Purportedly, C. parqui flowers in these North American locales from July through December. The fly specimens in this study were recovered from traps within this time frame, specifically 14 August to 9 November, 2013. Interestingly, several C. parqui plants were in flower at the University of California, Davis Arboretum in July and August, 2014, but Drosophila gentica was not found on the plants. In Chile, Cestrum parqui supports dense populations of Drosophila flavopilosa, a fly that occurs throughout southern South America [12, 13]. In addition, there are two native species of Cestrum in North America: Cestrum nocturnum L., native to the southern U.S. (California, Texas, Louisiana, Georgia, Florida), and which is planted as an ornamental in California, and Cestrum diurnum L. in Texas, Florida, and Puerto Rico [67]. The pharmacology of these species has been well investigated.

Our discovery of D. gentica in California indicates that the ecological niche modeling by [47] may need to be re-assessed. Using four flavopilosa-group species that had the best geographic sampling, and 19 environmental variables (analyzed using MAXENT), they determined that the fly distributions appear to be more limited by abiotic factors like temperature and rainfall than by the distribution of various Cestrum species. Unfortunately, sampling of flavopilosa-group flies is poor, both for hosts as well as geographically. Indeed, there are no published records of these flies for the entire Brazilian Amazon basin nor for the Orinoco Basin, including all of Venezuela and the Guianas, where Cestrum is known to also be diverse.

Drosophila flavohirta appears to be restricted to flowers of the large family Myrtaceae, which comprises more than 130 genera that are widely distributed in tropical and warm-temperate regions. To date, flies have been collected in large numbers only on flowers of Eucalyptus [25, 26, 36, 41] and Syzygium [26], and most recently found on Callistemon pallidus in Australia, all genera with massed, nectariferous flowers. The genus Eucalyptus contains c. 700 species, all native to Australia and nearby islands, various species of which have been introduced around the world, principally as a source of timber or shade trees since some species grow extremely tall and the resin (or “gum”, source of their popular name “gum trees”) makes them more resistant to insect damage [68]. Syzygium is a much larger and more widely distributed genus (c. 1200 species, with a center of diversity in southeast Asia); it too has been naturalized around the world, but not as extensively as Eucalyptus. Natural distribution of Callistemon s.s. includes the entire eastern margin of Australia and portions of southern Australia; this genus is closely related to Melaleuca [69].

Eucalyptus was introduced into California during the Gold Rush era of the 1850’s, and now there are some 250 species cultivated in the state [63, 68], with Eucalyptus globulus (blue gum) the most common species. Syzygium (as Eugenia) also occurs in California, but is much more limited in distribution and species within the state (some 44 species). Unfortunately, the flowers of Eucalyptus and Eugenia were not sampled during our study, only the flowers of the myrtaceous genus Melaleuca. Also, since the BioSCAN project is restricted to the Los Angeles area, we did not examine Malaise samples from other regions of California. Eventually, sampling should be made directly from myrtaceous flowers througout the state, especially the San Francisco area (an epicenter of California eucalypts), to determine how far D. flavohirta has spread. Given the diversity of Myrtaceae, especially in the Australasian Region, and knowing that Drosophila flavohirta is not restricted to Eucalyptus, it is possible that there are closely related species as yet uncollected in that region.

Several questions thus arise: When was D. flavohirta introduced into California? How far has it spread? Was Drosophila gentica introduced into California, or is it native? Does it (and perhaps other flavopilosa-group species) occur in other areas in the U.S. outside of California where Cestrum occurs? Native U.S. occurrence of D. gentica is highly doubtful, since it is extremely unlikely that such abundant flies escaped the detection of several large, very active Drosophila research groups in the region. In 1928 the famed “fly group” of Thomas Hunt Morgan moved to the California Institute of Technology in Pasadena [70], less than 20 km from most of our collection sites. This group included Alfred Sturtevant, author of the first treatment of North American Drosophila [71]. The other group was at the University of Texas, Austin, active from the 1940’s to 1970’s and where Marshall Wheeler was the systematist [13]. Drosophila flavohirta is much rarer in Malaise traps than D. gentica, somewhat expected since it congregates around arboreal blossoms and is rarely found near the ground. Drosophila flavohirta might also be rarely encountered if these are the early days of its establishment in California.