0 Introduction

Species discovery is the cornerstone for many aspects of biology ranging from species inventories，conservation to cataloguing the planets diversity. With an increasing pressure on natural habitats from a growing global population and threats from climate change，understanding the depth of biological diversity is an important prerequisite for devising effective conservation measures. With these adverse forces in mind，time is crucial in cataloguing the existing diversity.

On the other hand，increased interest in the natural world through education and initiatives，awareness of the natural diversity may lead to an increase in the number of new discovery of species. Such boosts in species discoveries may be noticeable by tracing the number of new species over time. Over the last few decades，a great number of plant species in the family Gesneriaceae Rich. & Juss. in DC. have been described from China. However，is there an acceleration in species discovery discernible?

The family Gesneriaceae has around 150 genera globally，with at present 71 genera in the predominantly Old World subfamily Didymocarpoideae Arn.^[1]. Forty three of these occur in China，spread across tribes Trichosporeae Nees (37 out of 64 genera)，and Epithemateae C.B.Clarke (5 out of 7 genera). A phylogenetic oddity is the genus Titanotrichum Soler. that occurs in Japan，the island of Taiwan and mainland China but has a new world phylogenetic association and is grouped in tribe Titanotricheae Yamaz. in subfamily Gesnerioideae Burnett^[2].

China is very rich and diverse in Gesneriaceae and harbours species from 9 out of 14 subtribes of subfamily Didymocarpoideae (and tribe Titanotricheae)^[2]. Xu et al.^[3-4] provided a recent account of the species in China where they focussed on the levels of endemism.

In the present paper，trends in species discovery rates are investigated，by plotting species discoveries against a timescale. Such study might indicate whether there is an accelerated rate of species discovery in recent years. Some comparisons of China are drawn with other areas where Gesneriaceae occur.

1 Material and Methods

Several databases were interrogated for baseline information for each species. These were primarily The International Plant Index (IPNI) (http://www.ipni.org/index.html，last accessed end of 2018) and the Smithsonian Gesneriaceae checklist (http://botany.si.edu/gesneriaceae/checklist/，last accessed October 2018)，amended with recent publications of protologues (see Table 1) and systematic realignments at genus level^{[2, 5-8]}.

The most recent delineations at genus level were used to define genera and cited in relevant sections^{[2, 9-11]}.For the total number of Gesneriaceae species published up to the end of 2018 of those genera，the publication year of the species' protologue was taken，irrespective of which genus the species was first described in. This was contrasted with only those species occurring in China. Varieties were not considered here. The label 'species' in the graphs was used for those species that never changed genus affiliation，i. e. remained in the genus they were first described in. The pattern of genus affiliation change per genus was traced and labelled 'renamed'.

The counts included came only from those genera that occured on China mainland and island of Taiwan. Since Cyrtandra J.R.Forst. & G.Forst. has over 800 species^[10] and only one occurs on the island of Taiwan，the species counts for this genus were omitted here.

2 Results and Discussion 2.1 Number of genera in China

The genera of Gesneriaceae underwent extensive restructuring from 2011 onwards resulting in new concepts for some for example Chirita Buch.-Ham.ex D. Don that was split into 5 genera^[9-11]. For those occurring in China the number changed from 56 to 43^[2]. Recently the genus Deinostigma W.T.Wang & Z.Y.Li，a monotypic genus with one species occurring in Vietnam was revised and expanded to include 7 species，two of which [D. cicatricosa (W.T.Wang) D.J.Middleton & Mich. Möller，and D. cyrtocarpa (D.Fang & L. Zeng) Mich. Möller & H.J.Atkins] occured in China^[6]. This brings the number of genera in China to 44.

2.2 Overall species numbers

The Gesneriaceae genera occured on mainland China and island Taiwan at the end of 2018 harbour 648 species (Table 1). This is an increase of around 65 species compared to an estimate made only two years ago^[2]. The difference in species number provided by Xu et al.^[3-4] of 671 and here is mainly due to their inclusion of the 55 varieties in China as 'species'，and the absence of 34 newly described species postpublication of their work in July 2017 (e.g. 15 species of Primulina，8 Oreocharis; see also Table 1).

Among those genera the most enlarged ones since Möller^{[2-3, 5-6, 12]} include Primulina Hance (+ 30 species) and Oreocharis Benth. (+ 14 species)，with others expanding by just under half a dozen species (Didymocarpus Wall.：3 species; Hemiboea C. B. Clarke：5 species; Petrocodon Hance：4 species; Petrocosmea Oliv.：5 species; Paraboea (C. B. Clarke) Ridl.：4 species) (Table 1). The increase did not necessarily reflect the species-richness of the genera，although for several larger ones with over 30 species in China，the increase in species number was with a 10% to 19% range，such as Didymocarpus (10% increase to now 33 species in China and 100 overall)，Hemiboea (16.1% increase to now 36 species in China and overall)，Paraboea (15.4% increase to now 30 species in China and 143 overall)，Petrocodon (14.3% increase to now 32 species in China and 34 overall) and Petrocosmea(14.7% increase to now 39 species in China and 47 overall).

Interestingly，a similar proportioned increase was found for the very large genera in China such as Oreocharis (13.7% increase to now 116 species in China and 125 overall) and Primulina (19.5% increase to now 184 species in China and 195 overall). On the other hand，for some genera，such as Aeschynanthus Jack^[13-14] and Epithema Blume^[15] taxonomic revisions resulted in reduced species numbers (Table 1).

2.3 Effect of genus delineation on species numbers over time

Taxonomic changes have an effect not only on genus size，but also on the species number over time. Not every species is necessarily described in the genus it resides in today. The taxonomic history has thus an effect on the pattern of species accumulation at the genus level. A few examples of distinct patterns are provided below.

In Petrocosmea，for example，very little taxonomic changes，i.e. two，were made since the genus was established in 1887. Hence，the line for Petrocosmea species number and protologues are virtually overlapping and the line plotting name changes is low and very flat (Fig. 1a). Its species occur predominantly in China，i.e. 39 out of 47，thus the Chinese species line is slightly below the one for total species (Fig. 1a).

Didymocarpus has a similar taxonomic history compared to Petrocosmea，but only a third of its species (33 out of 100 species) occur in China. Hence，the line for Chinese species is well below the line for the total Didymocarpus species number (Fig. 1b).

Oreocharis is a predominantly Chinese genus (Table 1). Until its delineation in 2011 it comprised 34 species. In 2011，species of another 10 Chinese genera were added bringing the species number to around 90^[7]. Since then，over 35 additional species have been newly described. For the above reasons，the lines for Oreocharis species and species in China overlap and are both well below the total Oreocharis species line. Because of the great expansion of the genus in 2011，the line tracing renamed species jumped at this point in time (Fig. 1c).

An example for another pattern is Primulina，also predominantly of Chinese origin (Table 1). However，until 2011 it was monotypic. In this year，it was enlarged to around 130 species，predominantly by the inclusion of species of Chirita section Gibbosaccus^[9-11]. This resulted in a very low and flat Primulina species line with a steep peak at 2011. Since then，additional species were newly described at a relatively steady rate of around a dozen per year to a total of 195 overall. Because of this history，the Primulina species line，Chinese species line and renaming lines overlap. The protologue line showed a steady increase in species from 1980 onwards，though these were described in other genera and later sunk into Primulina (Fig. 1d).

2.4 Rate of species discovery over time

The rate of species discovery over time showed some interesting patterns (Fig. 2). This showed a relatively constant increase over time with an average 5.9 species described per year across the 44 Old World Gesneriaceae genera，of which more than half，3.2，came from China alone (Table 2).

The chart in Fig. 3 showed a striking parallel development of species descriptions between all species and those occurring in China. It also showed some distinct accelerations in species discovery，which might be due to particular activities by taxonomists. For instance，a small peak in the 1879-1888 period could be traced back to the early monographic work by Clarke^[16]，though that was mainly covering species outside China. At this point in time，the rate of species description was 8.5 per year，with 3 species described from China per year (Table 2).

An even more prominent peak in the period between 1979-1988，which was to >90% due to work carried out in China that could be attributed to a large degree to the pioneering work of Wang Wen-Tsai (and co-workers) who lead the Gesneriaceae work for the Flora of China^[17]，with more than 200 names to his credit. During this period，a maximum of 41 species were described in 1983 from China，mostly by Wang Wen-Tsai，with an average of 17.4 per year for this decade (Fig. 3; Table 2). A third peak of alpha-taxonomic activity was discernible for the most recent decade，2009-2018，with around 2/3 (173) of the 271 new species coming from China (Figs 2，3; Table 2). A yearby-year breakdown indicated a relatively even spread across the decade with an average of 27.1 species per year overall and 17.3 from China (66.2%) (Table 3). Unlike the Wang Wen-Tsai era，here the species increase was based on the contribution of a wide range of botanists. This increase in botanical interest in the family resulted in the inclusion of no less than 67 different authors for species names over the last four years alone，with Fang Wen in particular highly prolific with an involvement in 25 species from China alone (Table 4).

2.5 Update on species number over last four years in China

The increase in species names in China over the last four years up to the end of the year 2018 was in total 80，relatively evenly spread between the years 2015 (23 species)，2016 (21 species)，2017 (22 species) and 2018 (14 species) (Table 4). All but seven were newly described taxa，while the two Deinostigma species and only species of Damrongia Kerr ex Craib，Deinostigma，Glabrella Mich. Möller & W. H. Chen and Middletonia C.Puglisi in China were recombinations，and Hemiboea pterocaulis (Z. Yu Li) Jie Huang，X.G.Xiang & Q.Zhang and Petrocosmea leiandra (W.T.Wang) Z.J.Qiu raised from variety to species status. Among the newly described species over this time period were 36 species in Primulina，15 in Oreocharis，6 in Paraboea，6 in Petrocodon，and 5 in Petrocosmea(Table 4).

3 Conclusions and Outlook

The present study investigated patterns of species discovery in China over time to investigate trends in the species description rate with the family Gesneriaceae as an example. With the addition of Deinostigma，currently species of 44 Gesneriaceae genera exist in China. The additional genus occurrence in China，however，was the result of the re-circumscription of a genus rather than species discovery. At the end of 2018，there were around 648 species (and around 55 varieties) of Gesneriaceae known in China，with many more in the process of being published. In fact in January 2019 there already was a new species published，Primulina anisocymosa F.Wen，Xin Hong & Z. J.Qiu^[18] (The publication date of 4 January 2018 in IPNI was seemingly an error). It was quite a challenge to keep a tally on the number of species in China since there was such an active community with > 60 botanists involved in the family Gesneriaceae. This was also reflected in the more than 30 species newly published species since the last review less than 2 years ago by Xu et al.^[3-4].

A note of caution should be inserted here with view to the high rate of new species emergence in recent years，and whether this indeed reflected the high species diversity of Gesneriaceae in China. The fastincreasing species number in this family may be affected by 'taxonomic inflation'，especially in the genus Primulina. For this genus，a species inflation has been strongly suggested in the complex around P. hochiensis (C.C.Huang & X.X.Chen) Mich. Möller & A.Weber^[19]. This can cause particular problems in macroecology and conservation^[20]. Approaches to circumvent this bias could include the application of population level sampling in molecular-based species delineation^[19]，rather than the often exercised inclusion of a single sample per new species，or to employ population-level sampling combined with morphological-molecular approaches^[21]. The often found approach to describe a single species per publication is a further unhelpful habit that，while inflating the publication record，scatters species information and a synthesis of the genus is difficult to obtain. Here，a proper revision at genus level particularly for the large and almost exclusive Chinese genera Primulina and Oreocharis would be most welcome to place the many species into context.

In the present survey it was found that the species were described with a relatively steady rate over time with about 3 species per year for China though with several peaks of activity in the 1980s when the average of new species per year increased to 17-18 species. Most of this could be attributed to the work of Wang Wen-Tsai. A further acceleration of species discovery had been revealed for the last ten years，with on average 17 new species per year. In this case however，it was an effort of over 60 different authors. As this trend seemed to persist to date，it might suggest that there would be many new species to be discovered in the near future，as it is unlikely that the entirety of China can be surveyed any time soon.

Nonetheless，the findings here should be an encouragement for botanists in China to keep up their floristic surveying，particularly in the light of increased pressures from agriculture and urban developments on natural habitats，and then，perhaps more difficult task，prevention of biodiversity loss in the light of global climate change. The wide basis of taxonomically interested botanists as illustrated in this survey is an excellent positive omen that this special expertise is on the rise in China. This skill is essential for inventorying the high diversity of natural diversity existing in China.

Acknowledgements

RBGE is supported by the Rural and Environment Science and Analytical Services Division (RESAS) in the Scottish Government.