Nature portrayed in images in Dutch Brazil: Tracing the sources of the plant woodcuts in the Historia Naturalis Brasiliae (1648)

By the mid-seventeenth century, images of natural elements that originated in Dutch Brazil circulated in Europe. These were often included in art collections (the Libri Picturati), and natural history treatises (the Historia Naturalis Brasiliae and the India Utriesque re Naturale et Medica). The plant woodcut images in these books constituted (icono) type specimens and played a significant role in the transmission of (botanical) knowledge. We present a systematic analysis of their origins by cross-referencing the visual and textual sources related to Dutch Brazil. To do so, we used the scientific identifications of the portrayed plants and digital archival material. The plant woodcuts accounted for 529 images, which correspond to 426 taxa. We created a PDF booklet to visualize the correlations of the woodcuts with the Libri Picturati and other visual sources. Substantial differences in the visual-making methodology exist between the two treatises (1648, 1658). Overall, availability, economy, and the Indigenous Tupi-based plant names that accompanied the images were crucial when arranging the sources, as well as portraying as much botanical information as possible. Freshly picked, living plants, and dried branches, fruits, and seeds were used to represent the megadiverse Brazilian flora, even when these belonged to species originating from other regions. Despite not being recognized for their contribution, Indigenous Brazilians and enslaved Africans were essential in the visual knowledge-making processes that later resulted in these natural history collections. As several sources remain lost and many histories yet untold, further archival studies and collaborative projects are pertinent to reveal the missing pieces of this conundrum.


100
bound collection (the Theatrum). Mentzel included Johan Maurits's words in the preface of 101 the Theatrum, emphasizing that 'these images aimed to reproduce Brazilian nature as 102 perfectly as possible' (10: 18). The elector was also gifted with two bound volumes of 103 watercolors (the Libri Principis), and a few sketches and oil paintings, which were bound in 104 1757 (the Misc. Cleyeri) (4). In the nineteenth century, these Brazilian collections were 105 incorporated into a larger collection known as the Libri Picturati, which was housed in the 106 Preussische Staatsbibliothek (the present-day Staatsbibliothek in Berlin, Germany), until this 107 library was evacuated during World War II (11). The Libri Picturati were considered lost 108 until zoologist Peter Whitehead located them at the Jagiellonian Library in Krakow, Poland 109 (4,12).  112 Due to its undeniable historical value, several studies on the Brazilian images in the Libri 113 Picturati have been conducted, focusing on its content, composition, and authorship 114 (4,20,21, 11,[13][14][15][16][17][18][19]. Scholars have argued that the Libri Picturati images served as models for 115 the woodcuts of the HNB (4,24, 25,11,12,[14][15][16]18,22,23). Correlations between the 116 Theatrum illustrations and the woodcuts had been established for the birds (26) and the 117 animals in general -also including the Libri Principis (15,25). A few woodcuts seem to be 118 based on Post's drawings (27). Brazilian botanists Pickel (6) and Almeida (5) suggested the 119 borrowing of images from published treatises, such as those by Clusius or De Laet. Some 120 woodcuts were made in the Dutch Republic after the specimens collected by Marcgrave (8). 121 Additionally, several woodcuts were based on Marcgrave's drawings, as De Laet indicated in 122 the preface of the HNB (1), as well as Piso in his book (2). In 1640, Marcgrave wrote to De 123 Laet from Brazil that he had made 350 pencil drawings of plants and several more of animals 6 124 (28). The few pencil sketches found in De Laet's manuscript could be part of those (4,8,22), 125 but most of Marcgrave's original drawings do not exist anymore or have not been located yet. 126 Beyond doubt, the study of the woodcuts embedded in these natural history books is of great 127 relevance, as they play a significant role in the transmission of knowledge (29) and they 128 constituted (icono) type specimens: they accompanied the first descriptions of individual 129 species against which all later individuals were compared (4). As the intriguing question of 130 which plant images were used as the basis for the woodcuts remained unanswered, we 131 present a systematic analysis of their origins. We focused on the botanical images included in 132 the HNB, the IURNM, the Libri Picturati Brazilian collection, De Laet's manuscript, and 133 Marcgrave's herbarium specimens to 1) analyze the correlations between the woodcuts and 134 the other visual sources from Dutch Brazil, and 2) trace back the remaining sources that were 135 used to create the woodcuts. By applying our botanical image analysis to these historical 136 collections, we provide an overview of how the visual material was used in the composition 137 of seventeenth century natural history treatises on Brazil. In doing so, we add insights into the 138 processes of visual knowledge-making and botanical practices in the early modern period.    because De Laet re-used them for the species that were mentioned by both authors (Fig 3). In 212 the HNB, a total of nine species are represented by two different woodcuts, while two are 213 depicted by three different ones. For six woodcuts we were unable to identify the species 214 depicted (Fig 3). 215 We traced approximately one-third (84 woodcuts) of the 243 unique woodcuts in the HNB 216 (Marcgrave's woodcuts plus Piso's unique ones) to the Libri Picturati, Marcgrave's 217 herbarium, or other known sources (Fig 3). Over two-thirds (159 woodcuts) of the plant 218 images did not correspond to any of these visual sources.      comparing the taxa shared between the HNB and the herbarium, we found that the woodcuts 254 of 25 species bear no resemblance to the vouchers; hence these were probably made after 255 other sources. For 17 species in the herbarium, we cannot infer whether the specimens were 256 used to design the woodcuts, because they are poorly preserved or consist of a few plant parts 257 (a single leaf or sterile branches). Nevertheless, 17 specimens bear similarities to the woodcut 258 images (Fig 3), so they could have been used as models to elaborate the drawings that were 259 later used to carve the woodcuts (Fig 6).   and portraits that are similar to the woodcuts in the HNB (S1, S2). One example is Ipomoea There is a crayon sketch of the same species in the Misc. Cleyeri (Fig 7e), but it neither 299 resembles the woodcut (Fig 7b) nor the vine in the painting (Fig 7d) -as is often the case 300 with the sketches in the Misc. Cleyeri and Eckhout's paintings (16,21).    that 'more details about this tree were to be found in Piso, to whom we owe this figure' (3:  In the IURNM, there are in total 228 woodcuts, eight of which are depicted twice (Fig 8). 337 Hence, 220 different woodblocks must have been used to complete the botanical part of 338 Piso's solo work. These are distributed in Piso's chapter IV, which is the equivalent to 339 Marcgrave's chapters on flora and the chapter on medicinal plants by Piso (44), and in Piso's 340 chapter V, which is the equivalent to the chapter on venoms and antidotes in the HNB (45).  Within the 59 identified taxa (Fig 8), we found a mushroom, a sponge, and a coral, which do 347 not belong to the plant kingdom but were classified as such by seventeenth century scholars.
Most of the woodcuts (72%) look exactly like those printed before in the HNB, suggesting 349 these were made with the same woodblocks that De Laet had used ten years earlier, and  the IURNM by two woodcuts (Fig 9a). One of them is the same as in the HNB (Fig 9b) and woodblock used for the HNB (Fig 9d), which was owned by their common publisher:   represented by a woodcut in the HNB (Fig 10a) of lesser quality and scientific detail than its 409 homolog in the Theatrum (Fig 10b). Despite the poor quality, De Laet included this image  (Figs 10c, 10d). This image was made 'au 419 naturel' after the seeds that De Laet (41: 137) received from Brazil and used in his treatise on 420 the Americas (Fig 10e).   (Fig 12a). In the same woodcut, we also see the inflorescence and the tiny (immature) 455 fruits, characteristic of the Capparaceae species (Fig. 12a). C. tapia in the Theatrum, 456 however, is represented by two illustrations glued in the same folio and each depicts a 457 fruiting and a flowering branch (including the fruit buds) (Fig 12b).     bears Icon in lib. 69 next to it (Fig 13a). Therefore, we would assume that the described plant 529 would include a woodcut in the HNB copied from a book or notebook numbered 69. The 530 description corresponds to Calea elongata Baker (6), but the woodcut in the HNB is very 531 similar (in reversed format) to the Theatrum illustration for Lantana camara L. (Fig 13b).

532
The oil painting bears the name Camaràuna (Fig 13c), which was presumably written by the 533 artist in question (Fig 13d). De Laet warned the reader in the HNB that 'the image we give 534 here [with C. elongata description], even though we found it under the name Camara uuba, 535 seems to be from another Camara, which the author [Marcgrave] mentioned before' (De 536 Laet's commentaries in (3: 6) (Fig 13e).  Since their images are only slightly similar, this weakens the possibility that they originated 551 from the same source.

552
Another species that has different local names in the HNB and the Theatrum, and is 553 represented with a different image, is Schinus terebinthifolia Raddi. A fruiting branch in the 554 HNB (Fig. 14a) is accompanied by the description of this species and the names Aroeira and   Ten years later, the woodblock must have been damaged or disappeared, as Piso (2) used a 572 slightly different image based again on De Laet's and Clusius' woodcuts (Figs 14b, 14c), 573 instead of the legitimate species, as the one present in the Theatrum (Fig 14d). This oil 574 painting includes the name Cambuí (Fig 14d), which we found in the HNB, but for a different 575 species: Eugenia involucrata DC., in (32: 82). Whoever wrote the reference above the 576 illustration of S. terebinthifolia in the Theatrum was pointing out the woodcut of E. 577 involucrata, but they noticed the discrepancy between the image and the description (Fig   578   14d).

579
The names Cambuí and Aroeira were used for S. terebinthifolia in seventeenth century Dutch  Arranging the visual and textual material to create the natural history treatises was an arduous 594 task. As the systematic binomial nomenclature had not yet been established, De Laet,595 Mentzel, and other scholars who used the HNB as a reference for tropical plants, such as 596 C.F.P. von Martius, struggled to organize the megadiverse Brazilian flora by their vernacular 597 names (65). They cross-referenced the different sources that were available to them, just as 598 we did in this study -although we used the scientific names first.

599
The difficulties of associating the visual material with the right vernaculars and their This material originated from a variety of people, who specialized in different subjects, but 644 whose skills were connected. Marcgrave was trained as a botanist and astronomer, but he also 645 made drawings and retrieved medicinal plant knowledge from the native population, just as 646 the physician Piso did. Marcgrave, however, was not only interested in the utilitarian value of 647 the flora he studied, hence some of his woodcuts belong to species that were not described as

666
Following Chen (76), we argue that the HNB created a new visual language by including 667 many images never seen before. Those images constituted a legitimate visual repertoire that 28 668 was later borrowed by others, such as Worm (56) or Piso (2), whose IURNM resulted in an 669 accumulation, rather than an innovation, of visual knowledge by replicating the images 670 published in the HNB a decade before. Piso also copied the woodcuts from other authors,  images served as models to elaborate the woodcuts of the HNB is answered in this research.

679
The Libri Principi's plant images were not used as the basis for any of the woodcuts in 680 Marcgrave and Piso's treatises. This differs from the animal woodcuts in the HNB, which

731
The repertoire of drawings that were used to elaborate the HNB and IURNM is incomplete.

732
Creating these treatises can be compared to making a puzzle with several pieces lacking, and