Tracing the introduction history of the tulip that went wild (Tulipa sylvestris) in sixteenth-century Europe

Tulipa sylvestris, commonly called the “wild tulip”, was introduced from the Mediterranean to northern Europe in the sixteenth century and became widely naturalized. Research has focused on tulips that came from the Ottoman Empire, but the introduction path of this native European, early ornamental tulip is unclear, and so is its taxonomic status: three subspecies are provisionally accepted, sometimes treated as species. Here we elucidate the history of introduction of T. sylvestris and discuss its taxonomy based on our historical findings. The first bulbs came from Bologna (northern Italy) and Montpellier (southern France) in the 1550–1570 s. Several renowned botanists were involved in their introduction, namely Gessner, Wieland, Aldrovandi, De Lobel, Clusius, and Dodoens. There were various introduction routes, including one from Spain which was apparently unsuccessful. The strong sixteenth-century Flemish botanical network facilitated the introduction and naturalization of T. sylvestris across Europe. Based on the latest tulip taxonomy, the diploid subspecies australis is native in the Mediterranean, and the tetraploid sylvestris is naturalized over Europe, but our historical findings show that both sylvestris and australis were introduced to northern Europe. This underlines the need to reconsider the taxonomic status of T. sylvestris, highlighting the importance of botanical history in understanding the complex taxonomy of naturalized cultivated plants.


Methods
Literature on tulips contains a plethora of scientific and non-scientific sources, the latter sometimes of questionable accuracy. We therefore focused our research on original sixteenth-century textual and visual evidence, including botanical publications, herbarium specimens, collections of drawings and historical archives. Tracing botanical species in pre-Linnaean texts is challenging, as a single plant name was often attributed to several botanical species, while a single species could have many different plant names. We thus started with compiling a list of "trusted" names for Tulipa sylvestris, which we retrieved from sixteenth-century herbarium specimens, the taxonomic identification of which we herein verified. For this we searched for specimens of Tulipa sylvestris in all surviving sixteenth-century herbaria 41 for which species lists have been published and/or images of the specimens are available. Using the names of the herbarium specimens as keywords, we screened the original works of major sixteenth-century botanists, among which Clusius, Dodoens, de Lobel, Gessner and also 25 editions of Matthioli's Commentaries on Dioscorides published between 1549 and 1600 in Italian, Latin, French and German. We also considered publications of the early seventeenth century by authors active during the sixteenth century, including Johann Bauhin's Historia Plantarum Universalis which was published by his son-in-law almost 40 years after Bauhin's death 42 . We further sought historical evidence of plant material exchange between southern and northern European sixteenth-century botanists in surviving archives, namely the Aldrovandian manuscripts kept at the University Library of Bologna, the digitized mail correspondence of Clusius (https:// clusi uscor respo ndence. huyge ns. knaw. nl/), and the Early modern physicians' letters of the German speaking area (www. aerzt ebrie fe. de). Names retrieved from textual, visual and archival sources were gradually added to the list of keywords in order to achieve maximal traceability. Regarding the Aldrovandian manuscripts, we consulted the digitized volumes at http:// moro. imss. fi. it/ aldro vandi/ and for the undigitized volumes, we screened the manuscript titles published by Frati et al. 43 and requested to the University library of Bologna images of the originals for all catalogues of plants that Aldrovandi sent to northern European naturalists during the second half of the sixteenth century.
To familiarize with the morphological diversity of Tulipa sylvestris sensu lato across Europe, we carried out observations on ca. 400 specimens of wild and cultivated plants from around Europe, kept at Naturalis Biodiversity Center in Leiden, the Netherlands (L). Images of these specimens can be viewed at https:// biopo rtal. natur alis. nl/, by using the advanced search and entering "Tulipa" under genus and one of the following terms under epithet: "sylvestris", "australis", "primulina", "biebersteiniana", "celsiana".
We created a map in Adobe Illustrator 26.0.3 depicting the original and naturalized distribution of Tulipa sylvestris in Europe and the range of the two subspecies based on floristic sources and distribution databases of individual countries where T. sylvestris occurs 26,27,29,35, . Precise occurrence within a country was not always available especially in eastern and southeastern European countries.

Results
Early illustrations, Wieland and Gessner, 1540-1550 s. The oldest surviving illustrations of Tulipa sylvestris are two watercolors contained in a two-volume handwritten manuscript from 1549, known as the Codex Kentmanus (Table 1). This manuscript was compiled by the German physician Johannes Kentmann (1518Kentmann ( -1574, who studied medicine in Padua, Venice and Bologna between 1547 and 1549. The first watercolor is named "Tulipa Turcica" and depicts a slender plant bearing a single flower with five tepals (Fig. 1a). This unrealistic representation-monocots do not have five tepals-is accompanied by Kentmann's uncertainty on the identity of this plant. In his annotations, probably compiled somewhat later (1550-1554), he wrote: "The Turks call this plant in their barbarous tongue 'Tulipa'; what it is I do not know" 68,69 . The epithet "Turcica" implies an Ottoman origin, but all images in the first volume of the Codex Kentmanus were drawn after plants that Kentmann observed in Italy 69 . The paper sheet on which this "Tulipa Turcica" was made carries a watermark depicting a triple mountain and cross, which originates from Padua 69 (see also 70 ), where Kentmann lived and drew several plants of the city's botanical garden 69 . During his two-year stay, Kentmann also travelled elsewhere in Italy, observing and drawing interesting and unknown plants 68,69 . The second watercolor of T. sylvestris in Kentmann's manuscript is named "νάρκισσος [nárkissos]; Lilionarcissus, Tulipae species", and depicts a robust double-flowered plant (Fig. 1b). It is included in the second volume of Codex Kentmanus, which only partly contains plants that Kentmann observed in Italy 69 . An Italian origin seems likely for this illustration as well, but it was unfortunately drawn on a sheet without watermark, nor is it accompanied by annotations. Therefore, we do not have actual evidence about its provenance.
Kentmann's illustrations reached the scholarly circles through the Swiss botanist Conrad Gessner (1516-1565), who borrowed Kentmann's manuscript between 1554 and 1555 10,69 . Besides publishing the first scientific description of a (red) tulip, Gessner 10 also wrote about Tulipa sylvestris based on Kentmann's "Tulipa Turcica" and acknowledged Kentmann as the source of origin for the name "Tulipa". In his notes about T. sylvestris, Kentmann 68 wrote that the name "Tulipa" is owed to the shape of the flower that refers to a Dalmatian cap ("pileoli Dalmatici"), an information that Gessner 10 and other authors later reproduced, e.g., [71][72][73] . Modern authors (e.g., 1 ) have argued that the name "Tulipa" derived from the Ottoman "turban" as a result of a misunderstanding, in which presumably de Busbecq was involved. But, although this information is plausible, we could not find it in sixteenth-century literature. Interestingly, towards the end of the century the English herbalist John Gerard (1545-1612) provided some insights here, writing that "after some days in bloom the tepal tips and margins in the flowers of T. sylvestris turn backwards resembling a Dalmatian or Turk's cap, called Tulipan, Tolepan, Turban, or Turfan, from which the plant took its name" 74 (Table 1).
Referring to Kentmann's "Tulipa Turcica", Gessner 10 provided a rough morphological description of T. sylvestris and, more interestingly, mentioned that he received seeds of this plant from Guilandinus Borussus, the latinized name of Melchior Wieland (ca. 1520-1589). Wieland was a Prussian botanist who took up his studies in Königsberg in 1544/45, continued them in Italy and settled in Padua, where in 1561, he was appointed head of the city's botanical garden, a position he held until his death 75,76 . When Gessner borrowed Kentmann's manuscript he kept copies of 142 illustrations of plants that caught his attention, among which also T. sylvestris 77 (Figs. 1a and 2a). Since Kentmann had already left Italy, Gessner asked Wieland to send him seeds of this plant from Padua. It was probably somewhere between ca. 1554 and 1559 that Wieland's seeds arrived in Zurich. We base this on circumstantial evidence, as in 1554 Gessner saw Kentmann's illustration and we know that he corresponded with Wieland already since at least 1556 (Gessner to Wieland, 3rd May, 1556, www. aerzt ebrie fe. de/ id/ 00045 404), and apparently must have hosted him in Zurich even earlier since Gessner in his letters called him his former "house guest" (hospes) 78 . In spring 1559, at the latest, Wieland left Padua to embark on a ca. two-year field trip to the Orient, as we know that in the summer of that year he had already travelled through Istanbul to Cairo (Wieland to Aldrovandi, 2nd June, 1559, www. aerzt ebrie fe. de/ id/ 00016 516). As far as we know, this was the first intended introduction of T. sylvestris northwards in Europe. Gessner must have carefully studied Kentmann's illustration over the years, as can be seen in his personal notes on the illustration (Fig. 2a). He noted the morphological difference of this tulip from the (famous) red tulip that he had observed in Augsburg in 1559, and that this was rather more similar to another plant that he had received from Ulrich Fugger in 1560. He also drew Wieland's seeds on the paper sheet and noted their morphological difference from another yellow tulip ("Tulipae luteae") that he had received from the French surgeon and collector Nicolaus Rassius (Rassé). This might have been a yellow-flowered variety of T. gesneriana from Rassius' garden in Paris.
Another important botanist of the early sixteenth century, the German Leonhart Fuchs (1501-1506), also had two illustrations of Tulipa sylvestris (Table 1), but their dating as of 1543-1547 and 1543-1548 79 is questionable. The plant layout in the illustration dated 1543-1547 (Fig. 2b) is remarkably similar to Kentmann's "Tulipa Turcica" from 1549 (Fig. 1a) www.nature.com/scientificreports/ Heinrich Füllmaurer 79 . It seems however unlikely that Fuchs, who never travelled to southern Europe, possessed already in the 1540 s a plant still unknown to northern Europeans (see also 70 on Fuchs' misleading identifications of Mediterranean plants). Moreover, Fuchs wrote that T. sylvestris was frequent in German gardens 79 , which again is inaccurate for this early period. This observation, however, might have been written later, or referred to daffodils in general.   www.nature.com/scientificreports/ One more early illustration of Tulipa sylvestris survives in another handwritten manuscript of Italian origin, I Cinque Libri di Piante, compiled by the Venetian patrician Pietro Antonio Michiel from about 1550 to his death in 1576 80 . Michiel's life-long masterpiece contains images of the plants that grew in his famous garden outside Venice and in the garden of Padua, which he curated between 1551 and 1555 80 . T. sylvestris appears in this manuscript as part of a group of yellow daffodils under the name "Narcisi gialli da volgari-Tulipa spetie" ("Yellow daffodils in the vernacular-Tulipa species") (Table 1, Fig. 3). Michiel wrote that this plant grew on the Santo Angelo mountain in Abruzzo and in Bologna. A catalogue of the plants cultivated in Michiel's garden surviving in the Aldrovandian manuscripts includes several daffodil species ("Narcissi varie spes"; 81 ), under which T. sylvestris is likely also meant.
Early herbarium specimens: Bologna, 1550 s. The oldest surviving specimens of Tulipa sylvestris also come from Italy, Bologna in specific ( Table 2). Two specimens are dated 1552 and 1553 and are included in the herbarium of Ulisse Aldrovandi (1522-1605) 82,83 , one of the most prominent naturalists of the sixteenth century, appointed professor of botany in Bologna from 1561 and for the subsequent almost 40 years. The herbarium of Aldrovandi contains a total of 18 tulip specimens, some of which originated from Bologna and others from Padua. Aldrovandi probably received the latter from Wieland, as evidenced from their surviving letters, which indicate exchange of plants and tulip bulbs between the two since at least 1554. The specimens of T. sylvestris both came from Bologna, the 1552 specimen representing a robust double-flowered plant (Fig. 4a), and the 1553 specimen a smaller, more slender, single-flowered plant (Fig. 4b). A third early specimen from Bologna is contained in the herbarium of Francesco Petrollini (Erbario B; [83][84][85], dated pre-1553 86 (Table 2).
Six more sixteenth-century specimens of Tulipa sylvestris survive today, four of which are also of Italian origin. One is contained in the En Tibi herbarium, also made by Petrollini around Bologna in ca. 1558 85 and another is part of an herbarium volume of unclear provenance (Erbario A), preserved together with Petrollini's Erbario B ( Table 2). The two combined are conventionally known as "Erbario Cibo" 84, 85 . Two more specimens are dated from 1563, the first collected by the German physician and traveler Leonhard Rauwolf somewhere in northern Italy 70 (Fig. 5a), and the latter included in the surviving herbarium of Andrea Cesalpino, made in Pisa for the bishop Alfonso Tornabuoni 87,88 (Fig. 5b, Table 2).
Two specimens of unknown geographical provenance are included in the herbarium of the Swiss botanist Felix Platter dated ca. 1552-1614, and a specimen of French provenance (Cevennes, Montpellier), dated most likely after 1600, is included in another Swiss herbarium, that of Caspar Bauhin from Basel (Fig. 5c, Table 2).   71 called T. syvlestris the "small tulip" ("Tulipa minor") and published a short morphological description and a woodcut that depicts two slender, single-flowered plants (Fig. 6a). Dodoens wrote that these small tulips grow somewhere in southern France. Florum was part of a series of popular illustrated botanical books published by the famous Antwerp-based publishing house Officina Plantiniana run by Christophe Plantin. The woodcut of Tulipa sylvestris published in Florum is a mirrored copy of a watercolor (Fig. 6b) contained in the Libri Picturati, a magnificent collection of over 1400 plant illustrations, which is Flemish in origin and presently kept in Krakow, Poland 89 . This watercolor is named "the small yellow tulip of Montpellier" ("Tulipa parva lutea. Monspel[liensis]"), a city located indeed in Table 2. Specimens of Tulipa sylvestris preserved in sixteenth-century herbaria. The indication "robust" or "slender" is an approximate estimation of flower size and leaf width. a The plant names of Aldrovandi's specimens were written at a later date. The uniform handwriting in this multivolume herbarium suggests that all plant names were written after the compilation of the last volume, so in 1586 the earliest. The mentioning of de Lobel's "Lilionarcissus Bononiensis" (published in 1576) in the two specimens supports this hypothesis. b Specimen consisting only of a flower, the rest plant parts are not present. c Origin and dating of specimen estimated herein based on information on the label: "minor Monspelio a D. Cherlero : major a D. Saltzman. ex horto Dei".    www.nature.com/scientificreports/ De Lobel used the name "Lilionarcissus Norbonensis" (Table 1), the epithet "Norbonensis" indicating Narbone as the geographical provenance of the plant, which corresponds to Provence and Languedoc, or generally southern France. "Lilionarcissus" was a name used for tulips in scholarly circles at that time 14 , but was superseded by the vernacular "Tulipa" that was finally established nomenclaturally by Linnaeus 20 . De Lobel probably dug out those bulbs between 1565 and 1567/8, as this period he was living in Montpellier for his medical studies and eagerly botanized the city's surroundings 95 .
De Lobel mentioned that he sent the bulbs to friends in Antwerp, but who were they? Perhaps the most wellknown naturalist of Antwerp was the apothecary Pieter van Coudenbergh, but unfortunately, the inventory of his famous garden 10 was published at least five years before de Lobel sent the bulbs. It is therefore not surprising that no plant in this catalogue could be matched to Tulipa sylvestris, the closest record being a reference to three daffodil species ("Narcissi tres species"; 10 ). Moreover, van Coudenbergh was not listed by de Lobel 96 among his close acquaintances, making it rather unlikely that he was the one who received the bulbs. More interestingly, de Lobel 96 mentioned in his Cruydtboeck two scribes from Antwerp who supported his work with their gardens. The first was Willem Martini, city scribe of Antwerp since 1565, known also for his involvement in the Dutch resistance against the Spaniards. The second was Jan van Hoboken, scribe of Antwerp at a later time, 1578-1590. It may be that any or both of these two men received de Lobel's bulbs, but further information about their botanical activity is not known.
Besides the two Antwerp scribes, de Lobel gave a long list of plant "facilitators" in which a certain Mr. Reynoutre stands out. This name is an alias of Saint Omer, first owner of the Libri Picturati, where the model watercolor of the Florum woodcut of Tulipa sylvestris is included. Saint Omer was a wealthy nobleman and owner of an estate and castle in Moerkerke, near Bruges in Belgium 90 . Florum's watercolor, which was later reproduced also in de Lobel's (and Clusius') books published by Plantin, was made in the same period (between ca. 1565 and 1568) that de Lobel dug out the bulbs in Montpellier, and this provenance is also mentioned in the watercolor's name ("Monspelliensis"). It may thus be assumed that de Lobel was also the direct or indirect source of the T. sylvestris bulbs that Saint Omer cultivated in his garden, the material based on which the species was for the first time described by Dodoens in 1568.

Origin Italy: Aldrovandi-Bologna and Apennines, 1570 s. Five years after de Lobel wrote about
the Narbone tulip, he introduced in literature the Bologna tulip, "Bononiensis Lilionarcissus luteus, sive Tulipa" 97 ( Fig. 7). He wrote that the Bologna tulip looks like the Narbone tulip "in leaf, stem and flower, but it is fragrant, significantly more vigorous and bigger", the latter a detail of taxonomic importance as will be discussed below. He also added that the Bologna tulip was sometimes double-flowered, as can be seen in herbarium specimens and illustrations of the time (Figs. 1b, 4a and 7). This time, de Lobel published also two woodcuts: one for the Narbone tulip, the same as Dodoens' woodcut from Florum, and one for the Bologna tulip, representing a doubleflowered individual (Fig. 7). Unlike the woodcut for the Narbone tulip, the model used for the woodcut of the Bologna tulip could not be traced among the Tulipa sylvestris illustrations and specimens that survive today. Its stylistic resemblance to Florum's woodcut points to another, perhaps lost watercolor that has not survived in the Libri Picturati collection presently kept in Krakow.
We do not know where de Lobel got his Bologna tulips from. They possibly grew since early in the city's gardens, as Johann Bauhin (1541-1614) reported having seen them in San Salvatore 42 , one of the oldest gardens in Bologna, existing at least since the early 1550 s 98 . Aldrovandi and Petrollini already possessed specimens (Table 2) before or shortly after their graduation in the early 1550 s 98 , so it must have been in the study material of medical  www.nature.com/scientificreports/ students at the university of Bologna. Nevertheless, this tulip is not easy to trace in the 1568-1582 inventory of the Bologna public garden, which Aldrovandi founded in 1568. Several yellow tulips are listed in this inventory, but they are all mentioned to have a Turkish origin (Aldrovandian manuscripts Ms. 002).
What we do know is that Clusius received Tulipa sylvestris directly from Aldrovandi, as Clusius personally recalled: "It grows in abundance in the Apennines from where Ulisse Aldrovandi, the Bolognese professor, dug it out and sent it to me from Bologna, many years ago" 14 . Clusius discarded de Lobel's "Lilionarcissus" in favor of "Tulipa" and used again the same woodcuts as Dodoens and de Lobel for the French and Italian plants 13,14 . Clusius also suggested a new place of origin for T. sylvestris in Italy: the Apennine mountains. He first used the epithet "Apenninea" complementary to "Bononiensis" ("Tulipa Apenninea sive Bononiensis"; 13 ) and later kept only the Apennine provenance, "Tulipa Apenninea"; 14 ) (Table 1), though still mentioned that he received the plant from Bologna.
In our attempt to trace when Clusius could have received the Apennine tulip from Aldrovandi, we found a catalogue dated ca. 1574-1575 99 , containing a list of plants that Aldrovandi sent to Clusius for the imperial garden in Vienna (Aldrovandian Manuscripts, Ms. 136/05, cc. 371-374). A certain "Tulipanum luteum" (yellow tulip) is listed therein, which may refer to Tulipa sylvestris and could be the plant that Clusius remembered having received from Aldrovandi. However, the possibility of a garden tulip variety with yellow flowers (T. gesneriana) cannot be excluded. What is certain is that in 1571, Clusius had received seeds of both Bologna and Montpellier tulips from one of his patrons, Jean de Brancion, a rich man from Mechelen, as Brancion himself declared in his letter to Clusius on August, 3rd that year (Table 3) 100 . Clusius further dispersed this material in his network. For example, in 1577 he sent bulbs of both tulips to his German friend Joachim Camerarius, instructing him how to care for these "Tulipas Bononieses" and "Mompelianas": "Do not mix them together with other tulips. Because they produce lateral bulbs and spread. It is better to put them in a distinct space […] so that they don't spread too far. Otherwise they would occupy the whole garden in a few years" 101 . Clusius 13,14 repeated this observation from his own garden, that these tulips produce lateral bulbs through stolons ("tenuibus nervis in latera", "lateral nerves at   (Table 3) 103 . The Bologna tulip was not only cultivated in northern European gardens but also in Italian ones. In 1606, the prefect of the Pisa garden, Franscesco Malocchi, sent to Clusius a list of the garden's most beautiful plants, among which the "Tulipa Bononiensis lutea odorata" ("scented yellow Bologna tulip") ( Table 3) 104 . Johann Bauhin reported having received the Bologna tulips in Montbeliard, N. France, from Guilhelmus Landgravius (probably William IV, Landgrave of Hesse-Kassel) and, like Clusius, he also reported the tendency of these tulips to multiply and spread 42 . The observations of these two rigorous botanists about the vegetative reproduction of T. sylvestris form the earliest evidence of naturalization for this species, a garden escape established today in the wild in many European countries.
Origin Spain: Aranjuez, 1580 s? Together with the Apennine and Narbone tulip, Clusius 14 grouped also another plant which grew on a hill close to Aranjuez, in central Spain. This "Tulipa Hispanica" was "similar to www.nature.com/scientificreports/ the Narbone tulip but slightly smaller". Clusius had not observed this plant during his Spanish travels, nor had he seen its flower, but got the information that the flowers are dark red outside. He recalled that this plant was introduced to Belgium by the gardener Francisco de Hollebeque, a distiller from Mechelen, who became gardener at the royal garden of King Philip II of Spain in Aranjuez in 1580 105 . So it was probably after that year that these Spanish tulips were brought to Belgium. Clusius 14 noted the difficulty to cultivate this tulip, whose bulbs only gave a single leaf and gradually perished in the Belgian environment.

History versus taxonomy. Connecting our historical findings with the current taxonomy and distribution
of Tulipa sylvestris, we get some interesting insights. Based on the species' current subspecific circumscription, the plants that are naturalized over Europe belong to the tetraploid subsp. sylvestris, while the plants that are native in the Mediterranean region belong to the diploid subsp. australis 11 . Subsp. sylvestris is apparently native also in (parts of) Italy 11 (Fig. 8). Our findings, however, show that actually both subspecies have been introduced to the north. From the overview of historical literature, we obtained several places of origin for T. sylvestris: the Cevennes mountains in southern France, Bologna and the Apennine mountains in northern Italy, and Aranjuez in central Spain (Fig. 9). Looking into the taxa that grow today in these regions, we see that only the plants that grow around Bologna belong to subsp. sylvestris. The tulips that grow in the Cevennes, Apennines and Aranjuez belong to subsp. australis 21,30,46,47,106,107 (Fig. 9). If any bulbs of T. sylvestris did come to Italy from the Ottoman empire, then these would have likely also been subsp. australis 44 (Fig. 8). Linnaeus 20 described Tulipa sylvestris as the small yellow tulip of Italy and France ("Tulipa minor lutea Italica" and "Tulipa minor lutea Gallica"). He copied this information from Bauhin's Pinax 19 , a monumental work that summarized plant names of sixteenth-century literature (Table 1). Our screening of original sixteenth-century sources revealed where exactly in Italy and France the first bulbs of T. sylvestris came from, when, and which botanists were involved in their introduction to northern Europe. The first material came from Italy in a still early period for botany as a field study 108 ; Between ca. 1554 and 1559, Wieland sent seeds from Padua to Zurich to Gessner. The latter was at that time the most famous naturalist beyond the Alps, and by sending him tulip seeds Wieland possibly hoped to increase his status in the botanical scene (which he achieved shortly after, in   14 reported having received T. sylvestris from Aldrovandi from Bologna, from material originating from the Apennine mountains, an event that possibly happened in the 1570 s. Plants with more reddish-tinted flowers were introduced from Spain 14 , but apparently did not flourish in northern European gardens. Totally different introduction paths are also possible. For instance, it has been suggested that Tulipa sylvestris, being a weed of vineyards, was unintentionally introduced northwards in Europe through the grapevine export, perhaps already since Roman times 1 . This path could have acted complementarily to the intended introduction of sixteenth-century naturalists. It is even possible that also bulbs of T. sylvestris were among the numerous kinds of tulips that arrived from the Ottoman empire to Italian gardens in the early 1550 s, as Kentmann's "Tulipa Turcica" and the 1568-1582 Bologna garden inventory imply. Whatever the routes of introduction of Tulipa sylvestris may have been, it was the strong sixteenth-century Flemish botanical network, supported by Clusius, de Lobel, Dodoens, Plantin and their rich friends with gardens, such as Saint Omer and de Brancion, which facilitated the introduction of T. sylvestris and its spread in Europe. This network was supplied with Mediterranean plant material from two main centers, southern France and northern Italy, which were popular botanical destinations for northern European naturalists 108 . Aldrovandi from Bologna was a direct source of such material 43 . Clusius had a large network of correspondents in Europe to whom he sent tulip bulbs 15 , so we may expect that he distributed bulbs of T. sylvestris to more correspondents than those reported herein. Archives and letters between sixteenth-century naturalists contain plenty of references to exchanges of bulbs, tulips and daffodils, under which T. sylvestris may also be meant. Unfortunately, in most cases these vague plant names could not be linked to botanical species. It was with de Lobel's works 93,97 that it became evident that a group of tulips, in fact T. sylvestris, grew wild on European (Mediterranean) ground. De Lobel's epithets "Bononiensis" and "Narbonensis" and Clusius' "Hispanica" clearly distinguished T. sylvestris from the exotic garden tulips that were imported from the Ottoman empire. The Bologna origin persisted in literature and almost a century after, T. sylvestris is explicitly indicated as the Bologna tulip in the 1657 inventory of the Bologna public garden ("Tulipa flore luteo minor Bononiensis", "the small yellow-flowered tulip of Bologna"), unambiguously distinguished from the several garden tulips of various colors and shapes ("Tulipanorum variae species, & variorum colorum") 111 . On the other hand, the evidence that has reached our days is dominated by the large archives of Clusius and Aldrovandi. If more information had survived about Wieland, Dodoens, de Lobel or other naturalists, we may have had another view of the introduction history of T. sylvestris.

Discussion
The sixteenth century was the time when the study of plants shifted from a focus on medicinal uses to an interest on plant diversity and taxonomy 85,112,113 . Tulipa sylvestris, and tulips in general, do not possess medicinal properties and were probably of no interest to classical authors and their successors. It is therefore no wonder that Pietro Andrea Matthioli, the most published botanical author in the sixteenth-century and faithful follower of Dioscorides, wrote nothing about tulips. The interest on plants as direct object of study was gradually shaped in the course of the century. It would be no exaggeration to say that this reformation took place in parallel with the introduction of tulips in Europe; these flowers of unprecedented beauty and no apparent use, other than ornamental. Matthioli did publish woodcuts of tulips, though, some of which imaginary, in his chapter about daffodils. Only after Matthioli's death, his opponents dared to publish their amended versions of Matthioli's Commentaries 112 . In such an occasion, Camerarius 114,115 suggested that a multi-branched daffodil of Matthioli may have actually been T. sylvestris, and he published a woodcut of the species representing a triple-flowered plant (Table 1).
Interestingly, de Lobel 97 had noticed that the Bologna tulip is bigger than the French one. One of Clusius' correspondents, Parduyn 103 , also noted the distinctiveness of the Bologna tulip (subsp. sylvestris) from the French and Spanish one (both subsp. australis): "Tulipa Bononiensis diversa ab Hispanica et Narbonensi" ("the Bologna tulip, which is different from the Spanish and Narbone tulip") ( Table 3). Indeed subsp. sylvestris has been suggested to be an autotetraploid derived from the diploid subsp. australis ( 11,30 , see also 106 ), and as such it is expected to be larger in size. On the other hand, subsp. sylvestris grows in richer soils than subsp. australis, another factor that favors a larger plant size. Female sterility was also suggested to affect plant size in subsp. sylvestris 116  www.nature.com/scientificreports/ Lobel also noticed that the Bologna tulip was fragrant and more vigorous ("vegetatiora") than the French one, which may imply that the former plant can easier establish in (and be favored for) a garden. He further wrote that Bologna tulips were sometimes double-flowered, a character which may also be connected to their tetraploid nature. Sixteenth-century plants from Italy were mostly visualized by double-flowered individuals (Fig. 1b, 4a  and 7), sometimes even triple (Table 1). Nevertheless, populations of subsp. sylvestris have mostly single-flowered plants, the division of the stem being rather a random, infrequent event. Apparently, the multiple flowers, like the pleasant scent, were a desired character for ornamental purposes. In fact, subsp. sylvestris may have been a naturalized cultivated plant also in Bologna, as it occurs in the city environs in secondary habitats, and its native occurrence has been questioned 21,106 . Already since the sixteenth and seventeenth century it was reported being cultivated in the gardens of Bologna and other Italian cities 42,104,111 . It could even be that the tetraploid Bologna tulip derived from material that was brought to the city's gardens from the Ottoman empire in the 1540-1550 s. The two subspecies have distinct ecological and altitudinal preferences and do not co-occur in the wild. The different forms illustrated in sixteenth-century literature, slender vs. robust, single-vs. multi-flowered plants (Tables 1 and 2, Figs. 1, 4, 5 and 7), may correspond to these two subspecies, the diploid australis and the tetraploid sylvestris, but this is not always straightforward. Since both subspecies were brought together in sixteenthcentury gardens, we would expect that triploids also occur around Europe, which is supported in literature 33,117 . The Apennine origin of Clusius 14 is somewhat enigmatic. Clusius lumped together the Apennine and Bologna tulips ( Table 1) [12][13][14] , but the first should belong to subsp. australis and the latter to subsp. sylvestris. One could assume that the Apennine origin was a misunderstanding of Clusius and that Italian plants came only from Bologna. Indeed, Clusius 14 wrote that the Narbone tulip looked like the Apennine tulip, but it was "generally smaller", a statement that seems confusing considering that the australis plants of the Cevennes would not be smaller than the (also australis) plants of the Apennines, but they would be smaller than the sylvestris Bologna plants. In physically examining the two specimens preserved in Aldrovandi's herbarium in Bologna, we noticed that only one of them (Fig. 4a) can be identified with certainty as subsp. sylvestris. The other specimen (Fig. 4b) can rather be identified as subsp. australis 25,32,33 , but depending on the identification source used it can also fall within the range of both australis and sylvestris 29 or even be sylvestris 118 . Clusius 14 also wrote that the flowers of the Narbone tulips were greener outside compared to the Apennine tulips. This statement adds to the confusion, considering that the flowers of the plants that grow (in abundance) in the Cevennes are reddish outside, and not greenish ( 29 , F. Hopkins personal communication). Plants with greenish outer tepals that belong to subsp. sylvestris are said to grow in the area, but are hard to find (F. Hopkins, personal communication). The Libri Picturati watercolor, which is probably a plant from the Cevennes, rather has greenish-red tepals (Fig. 6b). Actually, among sixteenth-century representatives of Tulipa sylvestris, only Clusius' "Tulipa Hispanica" was mentioned to have reddish color in the outer tepals 14 , which is a main discriminant character for subsp. australis vs. sylvestris in modern floras, including the Spanish one 47 .
The above make evident that the currently accepted distinction of subsp. sylvestris as a naturalized tetraploid in (northern) Europe and subsp. australis as a native diploid in the Mediterranean (up to Central Asia) does not conform with the history of introduction of Tulipa sylvestris, because both diploid and tetraploid plants have been introduced northwards to Europe. The introduced plants came from three or four different locations in Italy, France and Spain and are morphologically variable, although probably Spanish plants did not flourish after their introduction (Fig. 9). A new taxonomic assessment of Tulipa sylvestris is needed. This will be the next step of our research, employing genomics, morphometry and ploidy.
We suggest that a combination of botanical and historical research is necessary in order to understand the complex origin and taxonomic status of naturalized plants that have a past of introduction. Recent studies point in the same direction [119][120][121][122] . Large-scale initiatives of the last decades to digitize historical archives, books and herbaria, and make them publicly available online pave the way for such interdisciplinary studies 85,121,123 . Establishing global open-access databases with more historic material can provide new tools for research in botanical history 122 . Physical access to the original material remains essential for observing details, as conducted herein for example to verify the taxonomic identification of Aldrovandi's ca. 470-year-old herbarium specimens. Nevertheless, our focus in botanical history studies should be (remote) research on digitized images, as this enormously facilitates the study and accessibility of valuable treasures of our botanical heritage, without the risk of damaging them.

Data availability
Most of the data used in this article are included within the manuscript. All sixteenth-century books consulted are available online and links to their digitized content are provided in the list of references. Most of the illustrations and herbarium specimens used are available online at the websites of the libraries where they are kept. In order to get access to illustrations and specimens that are not available online, we requested copies to the libraries that hold the copyright; data availability in this case depends on the copyright holder policy.