Fig 1.
Maps of species sampling and the Ukaguru Mountains forested area.
A. Map of the Eastern Afromontane region showing the sampling localities of each member of the spiny-throated reed frog complex included in this study. Elevation is displayed from green colors near sea level and browns for higher elevation areas in the mountains (elevational range is 0–6,000 m above sea level). Mountain blocks and highland areas with spiny-throated reed frog populations are labelled. The locality of the new lineage described here, H. ukaguruensis sp. nov., is marked by a gold star, while all others are indicated by circles. All colors associated with each species correspond to the color scheme used in all figures. B: Elevational map and outlines of the Ukaguru Forest Reserves with locality of H. ukaguruensis sp. nov. marked with a gold star.
Table 1.
Summary information on sample sizes (number of individuals) for morphological and molecular datasets for each lineage.
The genetic dataset represents the number of individuals with contributed loci for analyses (16S, RAG1, POMC), with values for each in that order. Some individuals do not have available genetic information for all three loci included. All species except H. ruvuensis and H. ukwiva have genomic (ddRAD SNPs). New data generated in this study are all genomic data and new sequences for all genetic loci of two Sao Hill H. minutissimus and two H. ukaguruensis sp. nov. (indicated with asterixis). Breakdown of males and females for measurements are shown in Table 3.
Fig 2.
Phylogenetic relationships between members of the spiny-throated reed frog complex from both mitochondrial and nuclear markers.
The color scheme corresponding to each species is indicated on the phylogenetic tree and corresponds to all figures in the manuscript. A. Phylogenetic tree of specimens from the spiny-throated reed frog complex based on 16S mitochondrial data. Museum IDs or field IDs are listed for each individual included. Maximum clade credibility tree from BEAST2 shown, with branch lengths corresponding to sequence divergence. Values on nodes reflect posterior and bootstrap estimations from BEAST2/MrBayes/RAxML respectively. While the overall tree is mostly stable across estimation methods, the placement of Hyperolius tanneri is uncertain between methods, indicated by “-”and ** to show that this relationship is only supported in BEAST. B. TCS haplotype network based on the nuclear POMC alignment. Each circle node represents a unique haplotype and the size is proportional to the number of samples it represents. Only one haplotype has more than one species that share it, where one H. davenporti has a common H. burgessi haplotype. Species are indicated in color, and hypothesized intermediates are shown as solid circles. Crossed lines on the network indicate single nucleotide polymorphisms. C. TCS haplotype network based on the nuclear RAG1 alignment. All characteristics are shared with panel B. No haplotypes are shared between species and circles are not proportional to panel B. D. StarBeast species tree. All species are represented by at least one individual with all loci sequenced except H. ruvuensis which only has 16S data available.
Table 2.
Genetic distance (patristic) between species based on 16s sequence data.
Hyperolius minutissimus is broken down into two separate lineages based on locality (Sao Hill = SH, Uzungwa Scarp = US) due to significant structure. Standard error from 100 bootstrap replicates is shown in parentheses for each pairwise distance.
Fig 3.
Genomic comparisons between the members of the spiny-throated reed frogs.
A. Maximum clade credibility tree from BEAST2 shown, with branch lengths corresponding to sequence divergence. Values on nodes reflect posterior and bootstrap estimations from BEAST2, MrBayes, SNAPP, and RAxML respectively. Photos of each clade are shown. Hyperolius ukwiva is included as “inferred” due to its close mitochondrial and nuclear affinities in the genetic dataset to H. ukaguruensis sp. nov. and the results of the species tree (16S, RAG-1, POMC analysis). B. SNAPP tree showing gene trees from the top supported phylogeny (“blue trees”). This tree differs in whether H. tanneri is basal to all or in a basal sister lineage with H. ukaguruensis. C. STRUCTURE analysis showing K 3–5. Each species is indicated by the first letter of their name (e.g., H. burgessi–B, H. spinigularis = S, etc.).
Fig 4.
Morphological comparison of spiny-throated reed frogs.
A) PCA of all morphological variables comparing males of eight species. B) Snout-urosyle Length for females and males. C) Eye diameter/head width ratios for females and males. D) Gular flap width/height ratios for males. Letters above boxplots indicate post-hoc groups at p<0.001. The only known male specimen of H. ukwiva could not be located for measurements (only casual photos were available), thus measurements of this individual were not able to be included.
Table 3.
Mean and standard deviation of all measurements for H. spinigularis species groups.
Most data was taken from Lawson et al. (2015) and Barrat et al. (2017) with the addition of H. ukaguruensis data from this paper.
Fig 5.
Schematic drawings of the ventral view of head region of H. ukaguruensis sp. nov. and comparison to previous drawings from Barratt et al. (2017) and Loader et al. (2015).
Fig 6.
Dorsal and ventral view of anaesthetized male and female type specimens showing dorsal color polymorphism.
From left to right: BMNH 2022.7682 (holotype; male), field ID SL_4033 (male), BMNH 2022.7683 (female) and field ID SL-4057 (female). Photos by H. C. Liedtke.
Fig 7.
Photos in life of H. ukaguruensis sp. nov. A,B) Male (BMNH 2022.7682; holotype) and female (BMNH 2022.7683; paratype) in vivo, C) Hyperolius ukaguruensis sp. nov. male and female in axillary amplexus, and D) Type locality habitat. Photos by C. Liedtke.