Exploring the historical distribution of Dracaena cinnabari using ethnobotanical knowledge on Socotra Island, Yemen

Background In this study, we present and analyze toponyms referring to Socotra Island’s endemic dragon’s blood tree (Dracaena cinnabari) in four areas on the Socotra Archipelago UNESCO World Heritage site (Republic of Yemen). The motivation is the understanding of the past distribution of D. cinnabari trees which is an important part of conservation efforts by using ethnobotanical data. We assumed that dragon’s blood trees had a wider distribution on Socotra Island in the past. Methods This research was based on field surveys and interviews with the indigenous people. The place names (toponyms) were recorded in both Arabic and the indigenous Socotri language. We grouped all toponyms into five different categories according to the main descriptor: terrain, human, plant, water, and NA (unknown). Also, this study identified current and historical Arabic names of dragon’s blood trees of the genus Dracaena through literature review. Results A total of 301 toponyms were recorded from the four study areas in Socotra Island. Among names related to plants, we could attribute toponyms to nine different plants species, of which six toponyms referred to the D. cinnabari tree, representing 14.63% of the total phytotoponyms in the category. Three historical naming periods prior to 2000 could be identified. The most commonly used name for dragon’s blood trees (D. cinnabari, D. serrulata, D. ombet) appears to be “ahrieb” “إعريهب” and its resin “dum al-akhawin” “دم الأخوين,” while derived (mixed-cooked) products are called “eda’a” “إيدع,” while regionally different names can be found. Conclusion The place names that refer to D. cinnabari are herein suggested to represent remnant areas of once large populations. Therefore, the toponyms may support known hypotheses based on climate models that D. cinnabari had a wider distribution on Socotra Island in the past. This study also confirmed the historical importance of dragon’s blood.

137 Socotra and dragon's blood 138 The genus Dracaena is classified in the family Asparaga-139 ceae subfamily Nolinoideae (The Linnean Society of 140 London 2016). Dracaena cinnabari balf.f Socotra 141 dragon's blood tree belongs to the dragon's tree group 142 which contains 14 species as reviewed by Maděra et al. 143 [50] based on Marrero et al. [51] and Marrero [52]. The 144 Socotra dragon's blood tree (Dracaena cinnabari) is a 145 flagship species of Socotra [53,54]. It was a very import-146 ant tree in ancient times due to a historically highly 147 prized product called dragon's blood, a red resin ex-148 tracted for a wide range of uses including coloring and 149 local medicine [39,50]. Some believe that the name So-150 cotra could even be derived from "Sukkatira" or 151 contracted from "suq qatra," where suq is the Arabic 152 word for "market" and qatra for "dragon's blood," which 153 means "drop" related to the dropping of the liquid resin 154 pieces from the stem of the plant before drying it [55, 155 56]. The first who mentioned D. cinnabari resin was the 156 unknown author of the Periplus of the Erythrean Sea 157 around the mid-first century AD, who called it "cinnabar" 158 [57]. Dioscorides (90 AD) mentioned the resin in his 159 book "On Medical Material" as Kinnabari "cinnabari," 160 brought from Africa [58]. 161 Names of dragon's blood tree and its resin have been 162 recorded by old Arabic literature [59][60][61][62][63], by researchers 163 who visited Socotra [55,[64][65][66][67], and recently by [39, 46, 164 68, 69]. 165 Several local names for Dracaena may indicate the sig-166 nificance of the ethnobotanical knowledge as an import-167 ant source of information that can be used for tracking 168 the history of these names or link them to the land by 169 studying place names (toponyms). The aim of this study 170 is to use this ethnobotanical knowledge to explore the 171 spatial distribution of toponyms related to Dracaena cin-172 nabari tree and its potential as an information source to 173 assess the past distribution of this unique flagship spe-174 cies on Socotra Island. 175 Material and methods 176 Study area 177 Socotra Island is part of the unique Socotra Archipelago 178 natural UNESCO World Heritage Site (Republic of 179 Yemen), with a total area of 3,675 km 2 [70][71][72]. Live-180 stock grazing, fishing, agriculture (mainly date palm 181 plantation), and collection of non-timber forest products 182 are the main activities of people; the latter includes gath-183 ering resins such as frankincense, myrrh, and dragon's 184 blood and harvesting of aloe juice [39]. Socotra was fam-185 ous for these products in ancient times [39]   2). A variety of names for dragon's blood appear 245 during the golden era in science in the Islamic Arabic 246 world (ca. 800-ca. 1500 AD). The last period represents 247 the new western renaissance and scientific exploring 248 missions, especially from Europe. The described period 249 in this study begins in the year 60 AD, with the 250 appearance of the first name referred to dragon's blood,   3) were treated as one name because 266 they came from the same origin "cinnabar." Four common 267 names for the dragon's blood tree appear to be "dum al-268 akhawin," "eda'a," "al-huraifah," and "shian," besides the Eng-269 lish name "Dragon's Blood Tree," of course.

270
The contemporary names (Table   T3 3) were written accord-   Table   T4 4), which characterize how the t1:1   f3:1 Fig. 3 Frequency of appearing of dragon's blood (tree and resin) names in literature from 1st century AD to 2000 AD, T/R means that the name is f3:2 given to both the tree and the resin, references listed in 317 Toponyms related to animals and water have a lower 318 percentage of 9% and 6.9%, respectively, and there are 319 names of unknown meaning (10.6%).

320
Given the importance of plants for people in Socotra, it is not 321 surprising that the names referred to plants come in the third 322 position. Based on further analysis, we divided the plant names 323 into five subcategories (Table   T5 5) based on their meaning;   Greek name for red mineral (mercury sulfide HgS) and 365 adapted as the scientific name for Socotra dragon's f4:1 Fig. 4 Frequency of toponym categories in the four selected areas on Socotra, summary of 301 names f4:2 t4:1 akhwin as a common Arabic name for dragon's blood 370 resin, sometimes for the tree without specification of the 371 sources and for the resin brought from Socotra, and this 372 name continued to be used from the past until today. 373 "Eda'a" is the only local Socotri name that appears sig-374 nificantly in Arabic literature [60, 87] especially for the 375 resin of dragon's blood tree due to its famous use in 376 medicine. Other Arabic old names for dragon's blood 377 such as "andam" [76,83], "dum al-thuban" [76], and 378 "shian" [79] are also used for other products, as an ex-379 ample, "andam" used for logwood. "Dum al-tinnin" [76] 380 is the Arabic translation of dragon's blood, and "katir" is 381 the general name for drops. The first record for the local 382 name of dragon's blood tree was in 1899 by Forbes [55], 383 but eda'a and emsello ("moselle") have been mentioned 384 also by Wellsted [64]. Cabo González and Bustamante 385 Costa [93] suggested that there is a weakness in diction-386 aries and confusion of terminology related to dragon's 387 blood names and gave an example of "andam" and 388 "baqam." However, andam with its red color can bring 389 some confusion but baqam before 1500 AD was rarely 390 mentioned in Arabic literature, in my review just once 391 by Ibn Manzur and Mukarram [61]. From their reviews,

392
"shian" is a popular name for dragon's blood in Morocco 393 which is of Persian origin; however, the translation of 394 dragon's blood to Persian is "khun-siawshan" which ap-   (Table 2). In my opinion, 427 "Ahrieb" with its different form of writing and pronunci-428 ation is the common local name for all dragon's blood 429 tree species in Arabic region (Yemen,Socotra,Oman,430 and Saudia Arabia) (Table 3)  palm leaves such as "khwas" and "khazm;" the same is f6:1 Fig. 6 Frequency of phytotoponyms in the subcategory referring to identified plants (genus/species), including Dracaena cinnabari (27.3% of f6:2 the subcategory) f6:3 f7:1 Fig. 7 The map of the distribution of the toponyms related to the D. cinnabari tree (green circles) in Socotra Island, in red, the current f7:2 distribution of D. cinnabari by Maděra et al.
[54] f7:3 442 used for the leaves of dragon's blood tree in Socotra 443 "sa'af" which is also used for palm leaves. The names in 444 Morocco, Sudan, and Somalia are not linked to Arabic; 445 "ajgal" and "ombet" are in local languages and have been 446 used for the scientific name. 447 Toponym 448 The use of geographical-/ecological-based toponyms 449 stands as a potentially useful tool for aiding the recon-450 struction of historical changes. Toponyms have rarely 451 used as a biogeographic indicator of species or 452 vegetation-type occurrences [35]. Analyzing the topo-453 nyms (Fig. 5) shows a high frequency of names referring 454 to the terrain. Zeini et al.'s [94] study in Sinai (Egypt) 455 classified 69.9% of their recorded place names as refer-456 ring to the landforms followed by names that referred to 457 water. Human place names are typically metaphorical, 458 alluding to a resemblance between some physical feature 459 of a site and the shape of the organ after which it is 460 named [13]. In Socotra, human place names describe 461 where things happen and places where people harvest, 462 gather the goats, collect water, or play and they have a 463 general name for a whole area or landmark (like a 464 mountain). Plotting the distribution of plant names in 465 Socotra is another way to appreciate and display the eco-466 logical niche and knowledge.

467
Plant names (phytotoponyms), which are our focus in 468 this research, come in the third position with a fre-469 quency of 13.6%. Most of the phytotoponyms are for 470 general names or uses, while 53.7% of all phytotoponyms 471 could be identified by scientific names and 12.2% could 472 not be identified (Table 4). This shows a strong connection  All areas have a similar percentage of names referring to ani-491 mals that can be explained by the fact that grazing is com-492 mon in the selected areas and on the island in general.

493
Names with unknown meaning could be linked to ancient 494 language as suggested by Wagner (1960Wagner ( -1964  Dragon's blood tree name appears 6 times which repre-499 sent 14.6% of the phytotoponyms, four occurrences in f8:1 Fig. 8 Map of new sites of dragon's blood tree from the fieldwork (white polygon with dots), not published by Maděra et al. [54] (red color), and f8:2 possible sites for dragon's blood tree according to the local community for further field research (blue polygons), toponyms related to D. f8:3 cinnabari (green points) f8:4 500 western Socotra with two different variants and two in 501 eastern Socotra with also two different variants. Pérez 502 [95] also noted three different variants of the phytotopo-503 nyms "drago" in the Canary Islands, exploring the local 504 dialects. All the names in the western areas of the island 505 are not single names but linked with other words, D. 506 cinnabari pool, D. cinnabari sign, D. cinnabari place, 507 and D. cinnabari stand. The six names include two 508 names associated with the existing D. cinnabari tree, 509 one name associated with a place close to D. cinnabari 510 trees and where there have been trees before, one name 511 associated with a place close to D. cinnabari tree but no-512 body remembers that there was a tree before, one name 513 with no tree near but according to the local people it 514 could exist, and one name with no trees and historically 515 nobody knows of the trees existence on the area.

516
Half of the D. cinnabari place names occur in Qataria, 517 the area with a few limited isolated trees. Qataria is the 518 farthest western site of current D. cinnabari distribution 519 on the island, and the place names of D. cinnabari are 520 near and around the remnant D. cinnabari trees. Ac-521 cording to the local people in the area, there is a possi-522 bility of a small dragon's blood trees in the cliffs towards 523 the sea, which provides an opportunity for discovering 524 new Dracaena sites in the area. This area is the western 525 end of the ridge stretching from the central Hagher 526 Mts., and it indicates that the entire ridge could be likely 527 covered by Dracaena forests in the past, even though 528 Attorre et al. [47] did not assign most of this area to the 529 model of potential Dracaena cinnabari distribution. In 530 the second area in Momi, on the eastern side of the is-531 land, there is a D. cinnabari place name, but according 532 to the local community, there is also the possibility of D. 533 cinnabari trees on the cliffs towards the sea. In fact, this 534 area is not far from the recent Dracaena population and 535 was included in the potential Dracaena occurrence 536 made by Attorre et al. [47].

537
The third place is Ma'aleh (in the west), where D. cin-538 nabari is a place name but no more information is avail-539 able and there was no opportunity for visiting the site. 540 This finding is the most important because is the fur-541 thest from recent D. cinnabari distribution. In the entire 542 western part of Socotra, there is no one current record 543 of Dracaena occurrence beyond Qataria [54] and this 544 toponym would also confirm the model of potential 545 Dracaena occurrence in the Ma'aleh's highland pub-546 lished by Attorre et al. [47].

547
The last D. cinnabari place name in the Hagher is 548 close to D. cinnabari population, but the name is for a 549 place without D. cinnabari trees and they do not know 550 the presence of the tree in the place before; another D. 551 cinnabari place name was also recorded from Hagher 552 towards To'ahor's campsite-outside or study area (com-553 munication with local people).