Genomic evidence of a previously undetected Peromyscus truei invasion on San Clemente Island, California

Madeleine A. Becker123 , Kayce C. Bell4 , Selena Lopez-Ortiz2 , Jesús E. Maldonado2,3 ,
Cody W. Edwards1,3 , and Susette Castañeda-Rico1,2,3,* .

1Smithsonian-Mason School of Conservation, Front Royal, VA, USA. E-mail: beckerm@si.edu (MAB); cedward7@gmu.edu (CWE).

2Center for Conservation Genomics, Smithsonian National Zoo and Conservation Biology Institute, Washington, DC, USA. E-mail: lopezortizs@si.edu (SLO), maldonadoj@si.edu (JEM).

3College of Science, George Mason University, Fairfax, VA, USA.

4Natural History Museum of Los Angeles County, Los Angeles, CA, USA. E-mail: kbell@nhm.org (KCB).

*Corresponding author: susetteazul@gmail.com; castanedaricos@si.edu

Museum genomics (museomics) provide powerful tools for detecting cryptic diversity in natural history collections, including previously undetected invasions. We analyzed genomic and morphological data from 75 Peromyscus specimens collected during the Channel Islands Biological Survey (1939-1941), a series of expeditions on the California Channel Islands spearheaded by the Natural History Museum of Los Angeles County (LACM). Most of these specimens were confirmed as native Channel Island deer mice (P. gambelii subspp.), but two specimens from San Clemente Island (SCL) were re-identified as P. truei, a species never reported before on the Channel Islands. Using study skins, complete mitogenomes, and low coverage whole-genome data, we confirmed these identifications, and our phylogenetic analyses determined that these P. truei individuals are closely related to California mainland populations. We also placed these species within the subfamily Neotominae, including the first mitogenome of a native Channel Island deer mouse. Additional morphological data reveal several more likely P. truei specimens, indicating the existence of a substantial populations of this species in 1939. Combined with the presence of other non-native rodents on San Clemente Island (Reithrodontomys megalotis and Microtus californicus), these findings suggest a possible co-introduction from San Diego County during the 1930s via hay shipments. Our results have important implications for conservation management on San Clemente Island, and potentially Santa Rosa Island, where some specimens appear phenotypically inconsistent with P. gambelii. This study highlights the value of molecular tools for reassessing historic collections, especially in dynamic systems subject to high levels of anthropogenic modification.

Keywords: Channel Islands, cryptic diversity, island invasion, museomics, Peromyscus.

El estudio genómico de ejemplares de museo (museómica) es una herramienta poderosa para detectar diversidad críptica en colecciones de historia natural, incluidas invasiones previamente no detectadas. Obtuvimos datos genómicos y morfológicos de 75 ejemplares de Peromyscus colectados durante el Channel Islands Biological Survey (1939-1941), una serie de expediciones a las Islas del Canal de California encabezadas por el Natural History Museum de Los Angeles County (LACM). La mayoría de estos ejemplares fueron confirmados como ratones ciervo nativos de las Islas del Canal (P. gambelii subsp.), excepto dos ejemplares de la Isla San Clemente (SCL) que fueron reidentificados como P. truei, una especie nunca antes reportada para las Islas del Canal. El análisis de pieles de estudio, mitogenomas completos y datos del genoma completo de baja cobertura, permitió confirmar estas identificaciones. Los análisis filogenéticos mostraron que los ejemplares de P. truei están estrechamente relacionados con poblaciones continentales de California. Asimismo, incluimos estas especies dentro de la subfamilia Neotominae, incorporando el primer mitogenoma del ratón ciervo de las Islas del Canal. Datos morfológicos adicionales indicaron que varios ejemplares probablemente corresponden a P. truei, lo que sugiere la existencia de una población sustancial en 1939. Junto con la presencia de otros roedores no nativos en SCL (Reithrodontomys megalotis y Microtus californicus), estos hallazgos sugieren una posible co-introducción desde el Condado de San Diego durante la década de 1930 mediante envíos de pacas de heno. Nuestros resultados tienen implicaciones importantes para la gestión y conservación en SCL y, potencialmente en Isla Santa Rosa, donde algunos ejemplares parecen ser fenotípicamente inconsistentes con P. gambelii. Este estudio destaca el valor de las herramientas moleculares para reevaluar colecciones históricas, especialmente en sistemas dinámicos con altos niveles de modificación antropogénica.

Palabras clave: Diversidad críptica, invasión insular, Islas del Canal, museómica, Peromyscus.

Natural history collections (NHC) house irreplaceable records of global biodiversity spanning more than two centuries and increasingly provide the context needed to establish pre-perturbation baselines for threatened populations (Suarez and Tsutsui 2004; Benham and Bowie 2022). By coupling speci-mens with modern analytical tools, NHC enable retrospec-tive assessments of changes of geographic distribution, phenology, morphology, and genetic diversity over time, often driven by anthropogenic impacts (Meineke et al. 2019). Critically, museum genomics (museomics) of historical DNA can also pinpoint the origins and pathways of non-native taxa, resolving the number of introductions, their timing and spread (Kim et al. 2023), informing conservation responses to pathogens, disease vectors, and other biosecurity threats (e.g., O’Hanlon et al. 2018). As genetic barcoding and genomic tools are applied more broadly, NHC are revealing cryptic invasions events that have gone unrecognized in the natural history literature, and consequently, unaccounted for in conservation management decisions (e.g., Provan et al. 2008; Goedknegt et al. 2018).

The California Channel Islands, an isolated, near-shore archipelago, harbor exceptional endemic biodiversity and a rich archeological record reflecting over 13,000 years of human occupation on the northern islands and at least 9,500 years on the southern islands (Erlandson et al. 2011). Despite never being connected to the mainland, long-term human use, historical ranching, and 20th-century military activity have produced complex biogeographic legacies, including introduction of commensal and domestic animals and episodic translocations among islands and with the mainland.

Within this context, the Channel Islands Biological Survey (CIBS; 1939–1941) conducted the only comprehensive archipelago-wide survey to date, leading 13 expeditions and collecting thousands of specimens now curated at the Natural History Museum of Los Angeles County (LACM). San Clemente Island (SCL), the first island visited by CIBS, has experienced intensive sheep/goat ranching and US Navy activity, with documented ecological degradation followed by substantial eradication and restoration efforts (Keegan et al. 1994). These histories make CIBS holdings particularly valuable for reconstructing mid-20th century species composition and for detecting previously unrecognized introductions to the islands.

Island deer mice (Peromyscus gambelii sensu Greenbaum et al. 2017; 2019; Bradley et al. 2019) occur on all eight Channel Islands, each with morphologically and genetically distinct subspecies (Gill 1980; Ashley and Wills 1989; Pergams and Ashley 2002). Six island mouse subspecies had been described by the time of CIBS (Mearns 1897; Elliot 1903; Mearns 1907; Nelson and Goldman 1931), and CIBS mammalogist Jack von Bloeker would go on to describe the last two (von Bloeker Jr 1940; 1941). Their extensive Holocene history stretches at least 11,000 years, (Shirazi et al. 2018), likely driven by natural and human assisted dispersal, in addition to later contact with mainland lineages on the southern islands (Ashley and Wills 1989; Becker et al. 2025).

Recent introductions from other Peromyscus lineages have not been recorded on the Channel Islands, but given the archipelago’s dynamic anthropogenic footprint and the presence of other introduced rodents (von Bloeker Jr 1967; Rick 2013), they may be under-documented. Peromyscus are a diverse group of morphologically similar animals which have been difficult to resolve with traditional tools (Platt et al. 2015; Castañeda-Rico et al. 2025), and contain frequently misidentified lineages (e.g., Light et al. 2021). Furthermore, field identification within Peromyscus relies on the geographic and ecological context of collection—information which is no longer useful in novel insular contexts. Lastly, rapid morphological adaptation to novel island habitats is common in rodents (e.g., Pergams et al. 2015), further impeding accurate identification of newly established populations and raising the possibility of cryptic introductions.

While conducting population genomic research with Channel Island deer mouse specimens (Becker in prep.), we recovered genomic data from two individuals on SCL (Figure 1) that were inconsistent with identification as P. gambelii. Here, we revisit the CIBS collections with museomics and morphology to interrogate this finding and reassess CIBS Peromyscus specimens. We analyze mitogenomes, low coverage whole-genome data, and study skins and skulls from 75 specimens collected during the CIBS to: 1) test for non-native Peromyscus lineages within historic Channel Island collections; 2) phylogenetically place newly identified lineages; 3) evaluate introduction pathways and management implications. We report the first evidence of P. truei on the Channel Islands, identify additional likely P. truei specimens on SCL consistent with an invasion of multiple individuals, and infer a probable co-introduction pathway via hay bales shipped from San Diego County on the California mainland. Our study demonstrates the power of genetic identification in dynamic systems such as the Channel Islands and underscores the importance of revisiting historical collections with genomic tools.

Materials and methods

Specimen subsampling. We subsampled 75 Peromyscus study skins housed in Natural History Museum of Los Angeles County (LACM) mammal collections (Supplemental Data SD1A), primarily to study historical P. gambelii population genomics (Becker in prep.). Using sterile techniques, we performed whole claw excision due to documented high performance for this sample type in rodent study skins (McDonough et al. 2018). These specimens were originally collected across the eight California Channel Islands during the Channel Islands Biological Survey from 1939-1941. All specimens were catalogued as Peromyscus maniculatus, with original specific locality data matching each corresponding island subspecies. Specimens were not morphologically examined during subsampling, since native island mice are the only extant recorded Peromyscus occurring on any of the Channel Islands. Ten specimens were sampled from San Clemente Island, including two that we are presently re-identifying as P. truei based on morphological and genomic evidence (Figure 2).

Laboratory processing. Specimen subsamples were processed in a specialized ancient DNA facility with no exposure to PCR products or modern DNA. All laboratory work was conducted at the Center for Conservation Genomics (CCG), Smithsonian National Zoo and Conser-vation Biology Institute. We extracted samples with a modified ancient DNA silica-column protocol (McDonough et al. 2018), using either MinElute columns (Qiagen Inc., Valencia, CA, USA) or Zymo Spin Columns (Zymo Research, Irvine, CA, USA). DNA was quantified with 1x dsDNA High Sensitivity Qubit assays (Thermo Fisher, Waltham, MA, USA), and genomic libraries were prepared with the Santa Cruz Reaction customized for each sample’s input quantity (Kapp et al. 2021). Libraries were indexed with KAPA uracil-tolerant polymerase (Roche Kapa Biosystems, Wilmington, MA, USA) in a modern DNA facility, using Tru-seq style indices. Indexing PCR cycle number was determined by performing quantitative PCR on diluted pre-PCR libraries (Kapp et al. 2021). In addition to these specimens, we also prepared genomic libraries with the same methods from a P. truei truei tissue sample (USNM 603235, coll. 2017) sheared DNA extract, which was previously sequenced for captured ultraconserved elements and the mitochondrial genome (Castañeda-Rico et al. 2025). Genomic libraries were quantified with Qubit and Tapestation High Sensitivity (Agilent Technologies, Santa Clara, CA, USA), equimolarly pooled with other projects, and quality controlled for insert size with Blue Pippin (Sage Science Inc., Beverly, MA, USA). Final pools were submitted for low-coverage paired-end sequencing (150bp PE, 300 cycles) on an Illumina Nova-Seq X 25B flow cell (Illumina Inc., San Diego, CA, USA) at the Oklahoma Medical Research Foundation’s Clinical Genomics Center.

Mitochondrial genome assembly and initial species identification. Raw sequences were trimmed for adapter content, quality, and length with default parameters in Trim Galore v0.6.10 (Martin 2011), and exact duplicates were removed with prinseq-lite.pl v0.20.4 (Schmieder and Edwards 2011). Mitochondrial genomes (mitogenomes) were assembled from the processed shotgun sequencing data in Geneious Prime® 2025.0.3 using the Geneious mapping algorithm with default parameters and up to five iterations. Initially, all sequences were mapped to the Peromyscus maniculatus bairdii reference mitogenome NC_039921, but final versions of the LACM 5833 and LACM 5834 P. truei mitogenomes were mapped to the reference P. truei mitogenome PP818778 (USNM 603235) for increased mapping quality. Any sites under 5X depth and all ambiguous bases in the consensus sequences for these mitogenomes were called as N. Annotations were transferred from the reference mitogenomes in Geneious and checked for premature stop codons and the presence of nuclear copies of mitochondrial DNA (NUMTs). Cytochrome b gene sequences were extracted from mitogenome assemblies for species identification and/or downstream analysis, since only one P. truei mitogenome (PP818778) and no island mouse (P. gambelii spp.) mitogenomes have previously been published. Initial species identifications were made by querying NCBI’s core nucleotide database (core_nt) with cytochrome b sequences using BLASTN v2.17.0+ (Camacho et al. 2009). Of the confirmed island mouse specimens, we only included data from one SCL specimen (LACM 6722) in further molecular analyses to investigate differences between the P. truei collected with native Peromyscus on this island.

Nuclear DNA species identification. Since we obtained shotgun sequencing data from LACM 5833 and LACM 5834 as well as the specimen associated with the P. truei reference mitogenome (USNM 603235), we wanted to ensure that nuclear DNA also supported the species identification obtained from mitochondrial DNA. To do this, we used Kraken2 v2.1.3 (Wood and Salzberg 2014) to construct a custom database containing all available reference genomes within the Neotominae subfamily: P. maniculatus. bairdii (HU_Pman_2.1), P. m. sonoriensis (mPerMan1.0.p), P. leucopus (UCI_PerLeu_2.1), P. eremicus (PerEre_H2_v1), P. californicus (ASM782708v3), P. polionotus (HU_Ppol_2), P. melanophrys (Pmel_10x_v1), P. attwateri (Patt_10x_v1), P. nudipes (Pnud_10x_v1), P. aztecus (Pazt_10x_v1), Onychomys torridus (mOncTor1.1), O. arenicola (OncAre_H1_v1), O. leucogaster (OncLeu_H2_v1), Reithrodontomys megalotis (mReiMeg1.0.p), Neotoma floridana (mNeoFlo1.hap1), N. lepida (ASM167557v1), Scotinomys teguina (ASM4990163v1). Since shotgun sequencing of museum specimens also frequently includes common contaminants, we also added the pre-made “bacteria”, “human”, and “UniVec_Core” libraries to prevent false hits. Because a genome for P. truei has not yet been assembled, we added the quality-checked, merged reads of the USNM 603235 tissue sample as representative sequences of this species. All validated reads from LACM 5833, 5834, and 6722 were queried against this custom database, and krakentools v1.2 (Lu et al. 2022) and krona v2.8.1 (Ondov et al. 2011) were used to visualize the results.

Whole mitogenome analysis and molecular dating. We conducted multiple whole mitogenome phylogenetic analyses with the three novel mitogenomes and previously published data to 1) robustly confirm the initial P. truei species identification of LACM 5833 and LACM 5834, 2) resolve the relationship between these specimens’ mitochondrial haplotypes and closely related taxa, 3) and establish the position of P. gambelii clementis (LACM 6722).

We used MAFFT v7.490 plugin (Katoh and Standley 2013) in Geneious to align complete mitogenomes generated in this study and in Castañeda-Rico et al. (2025). Phylogenetic trees were obtained using Maximum Likelihood (ML) in IQTree v2.3.1 (Minh et al. 2020) and Bayesian Inference (BI) in MrBayes v3.2.7a (Ronquist and Huelsenbeck 2003) on partitioned mitogenome alignments. We used ModelFinder, as implemented in IQTree, to find the best-fit model for the ML analysis. IQTree was run with the best-fit models and partitions (TIM2+F+R4, GTR+F+I+R4, GTR+F+I+R5, GTR+F+I+R6, GTR+F+R4; five partitions) and 1000 ultrafast bootstrap replicates, with values of >95 indicating support. We used PartitionFinder 2.1.1 (Lanfear et al. 2017) with linked, corrected Akaike Information Criterion (AICc), and greedy parameters, to select the best model and partition scheme for the alignment. The search was limited to the models available in MrBayes. We defined the data block by codon position, tRNA, rRNA and D-loop selection. MrBayes was run with the best-fit model and partitions (GTR+I+G for each of the six partitions) for 50 million generations, with a burn-in of 25%, a sample frequency of 1000 generations, and 4 Markov chains Monte Carlo (MCMC). Posterior probabilities and final topologies result from 50% majority rule consensus trees, with a probability of >0.95 indicating support.

We estimated divergence times in BEAST2 v2.6.7 (Bouckaert et al. 2019) using a partitioned alignment. The best-fit model and partition scheme were selected with PartitionFinder 2.1.1, restricted to models available in BEAST and using linked, AICc, and greedy parameters. The data block was defined by codon position, tRNA, rRNA and D-loop selection. We ran the analysis with the best-fit models and partitions (GTR+I+G+X, TRN+I+G; six partitions) under an uncorrelated lognormal relaxed molecular clock model, a calibrated Yule speciation processes model, and with a randomly generated starting tree as priors. Calibrations were based on fossil records of 1) Onychomys sp. at ca. 4.85 million years ago (mya), 2) Peromyscus sp. at ca. 4.85 mya, and 3) Reithrodontomys wetmori at ca. 4.5 mya (see Castañeda-Rico et al. 2025). We performed two independent runs of 50 million iterations each, with sample frequency of 1000 iterations. Tracer v1.7.1 (Rambaut et al. 2018) was used to check convergence for the Effective Sample Size (ESS), and 25% burn-in was performed in each run. LogCombiner v2.6.6 and Tree Annotator v2.6.2, both packages in BEAST, were used to combine trees and to compile a maximum clade credibility tree with node heigh distribution.

All analyses were performed on the Smithsonian Institution High Performance Computing Cluster (Smith-sonian Institution, https://doi.org/10.25572/SIHPC). Trees were visualized with ggtree (Yu et al. 2017) in R version R v.4.5.1 (R Core Team 2021).

Phylogeographic analysis of Peromyscus truei with cytochrome b. We conducted phylogeographic analysis with extracted cytochrome b genes of LACM 5833 and LACM 5834 and previously published Peromyscus truei sequences to infer the origins of the introduction of P. truei to SCL. We downloaded all sequenced cytochrome b gene fragments greater than 1000 bp from NCBI Genbank (n=82), which were all sequenced from wild-caught museum specimens with locality information. We only analyzed unique haplotypes (n=67) in addition to LACM 5833, LACM 5834, and the P. truei reference PP818778 (USNM 603235). We chose related outgroups based on the Neotominae mitogenome phylogeny (Castañeda-Rico et al. 2025): P. gratus gratus (AY376421), P. gratus gentilis (AY376420), P. gratus zapotecae (AY376423), P. ochraventer (PP818776), P. attwateri (ON528112), P. nasutus (PP818775). P. pectoralis (KY707309), and Habromys ixtlani (KY707304). The resulting dataset (n=78) was aligned with the MAFFT v7.490 plugin in Geneious.

Phylogenetic trees were built using ML (IQTree v2.3.1) and BI (MrBayes v3.2.7a) on unpartitioned cytochrome b alignments. We also used ModelFinder as implemented within IQTree to find the best models for both analyses. IQTree was run with the best-fit model (TIM2+F+G4) and 1000 ultrafast bootstraps, with a score of >95 indicating support. We ran MrBayes with the best-fit model (GTR+F+G4) for 20 million generations with a burn-in of 25%, a sample frequency of 1000, and 4 MCMC. Posterior probabilities and final topologies result from 50% majority rule consensus trees, with a probability of >0.95 indicating support.

Morphological analysis. We examined study skins of the SCL specimens included in our genomic analyses (Figure 2) and compared their original measurements to those of genetically confirmed P. gambelii and P. truei to other Peromyscus specimens collected during the CIBS across the Channel Islands. The goals of the morphological analyses were: 1) to determine whether LACM 5833 and LACM 5834 skins comported with morphological characters consistent with the genetic identification of P. truei, and 2) to detect any other putative misidentifications present in LACM Channel Island deer mice collections. We also examined a subset of skulls to assess dental traits diagnostic of P. truei, specifically the presence of accessory cusps (mesolophs) on upper molars (Ingles 1965).

Of the total 709 Peromyscus specimens collected during the CIBS and housed at LACM, we restricted analyses to the 703 specimens prepared by Jack von Bloeker to ensure measurement consistency. Juveniles and subadults were retained due to inconsistent age class labels, but two extreme outliers in tail or total length (due to tail damage) were removed, resulting in a final dataset of n=701 (Supplemental Data SD1B). This dataset includes 277 specimens caught on SCL, including those sequenced for genomic analysis. Since P. truei typically exhibits larger body size with longer tails proportional to their body size and larger ears compared to P. gambelii/maniculatus (Ingles 1965), we plotted the distribution of these traits using von Bloeker’s original measurement data. Each island was treated as a separate population, and genetically confirmed individuals were concurrently plotted to visualize where P. gambelii and P. truei fell within island-specific distributions. These analyses were not performed to compare P. gambelii and P. truei specimens, but to better understand how our genetically identified individuals compared to the morphological variation found in each island Peromyscus population. To assess potential confounding effects of sexual dimorphism, pregnancy, and immaturity, we also plotted trait distributions by sex, known reproductive status, and recorded age class. All calculations and plots were made in R v.4.5.1 with tidyverse packages (Wickham et al. 2019).

Six specimens collected on SCL during the CIBS and now housed at the Santa Barbara Museum of Natural History were also examined, but were ultimately not included in this dataset as three of these specimens were originally identified as Onychomys and mistakenly re-labeled as P. maniculatus. We also examined 157 Channel Island Peromyscus specimens housed at the Smithsonian National Museum of Natural History (USNM) and collected in 1892-1931, prior to the CIBS. These specimens were largely collected before mammal body measurements were standardized, with many incomplete datapoints including few ear measurements. They were also prepared by several different collectors at multiple time points, so they were not included in the CIBS dataset. However, these collections remain an important reference point, and were still checked for diagnostic P. truei external characters such as high tail to body size ratios and tail tip fur density (Ingles 1965).

Results

Mitochondrial genome assembly and initial species iden-tification. Mitogenomes from LACM 5833, 5834, and 6722 were successfully recovered from assembled shotgun data, with average depths of 70-104.4X (Supplemental Data SD2). However, likely structural variation in the D-loop led to low depth and poor assembly at the end of this region for the two P. truei samples. Reads were re-assembled with GetOrganelle to try to recover mitochondrial genomes without these small ambiguous regions, but this method resulted in 3-6 short contigs and did not cover the entire mitochondrial genome. Nonetheless, the final Geneious assemblies only contained 34-51 ambiguities, and variants in the rest of the mitochondrial genome were sufficient for downstream analysis. Annotated mitogenomes for these three specimens have been deposited in GenBank (PZ213495-PZ213497), and raw reads for these specimens and USNM 603235 can be found on the Sequence Read Archive (SRR37837849-SRR37837852; Supplemental Data SD1C).

The mitogenomes of LACM 5833 and LACM 5834 differed by 159 nucleotide positions, corresponding to a pairwise identity of 99.0%. Both haplotypes returned the same top BLAST hit: the P. truei reference mitochondrial genome (GenBank accession PP818778; voucher USNM 603235), with 96.04-96.22% pairwise identity. In contrast, the SCL island deer mouse (LACM 6722) matched most closely with a P. maniculatus rufinus voucher (USNM 603271, Figure 3), with a 96.25% pairwise identity. Pairwise identity between the P. truei collected on SCL and the P. gambelii clementis (LACM 6722) mitogenome was 86.5-86.7%.

The cytochrome b sequences extracted from LACM 5833 and 5834 were also unique, differing by nine nucleotide positions across a 1,144 bp region. Assemblies in this region were of high quality, with an average sequencing depth of 83-104X and a minimum depth of 53-55X. Neither haplotype had an identical match in the NCBI core_nt BLAST database; however, both showed high percent identity to multiple available vouchered P. truei specimens from California. The top BLAST hit for LACM 5833 was MVZ:Mamm:198610, a Peromyscus truei montipinoris collected in Kern County, CA (99.56%), with 30 total hits over 99% percent identity. For LACM 5834, the top BLAST hits were MVZ:Mamm:157332 and MVZ:Mamm:157330 (Peromyscus truei gilberti) collected in Berkeley, CA with 99.48% identity and 10 total hits over 99% percent identity.

The cytochrome b sequence extracted from LACM 6722 was 99.83% identical to previously sequenced island deer mice from San Clemente and Santa Catalina Islands (Becker et al. 2025). All other LACM Peromyscus subsampled from the CIBS expedition harbored either identical or very closely related haplotypes to previously reported Channel Island deer mouse cytochrome b sequences. Nine novel haplotypes were identified and deposited on GenBank (PZ213486-PZ213494; Supplemental Data SD1C). These sequences were generated primarily to confirm species identification in morphological re-examination of specimens; no further molecular analysis were conducted on these samples.

Nuclear DNA identification. Validated shotgun reads from LACM 5833, 5834, and 6722 all contained high amounts of cricetid DNA (81-87%) according to the Kraken2 custom database search. Each library contained small numbers of human and bacterial reads and a significant number of unclassified reads (13-19%). This contamination is consistent with the age and preservation of the specimens (coll. 1939), and the unclassified reads may likely include fungi or other eukaryotic microbes found on study skins or pests present in collections (Raxworthy and Smith 2021). The highest species-resolved hit for LACM 5833 and 5834 was P. truei, with 31% of reads classified to this species. Despite our biased method of adding P. truei shotgun sequencing data to the custom database versus using an assembled genome like the other rodents, 22% of P. gambelii clementis reads were successfully classified as most related taxon of Peromyscus maniculatus, with a higher percentage of reads only classified to the genus level (43%). 1% of LACM 6722 reads were classified as P. truei, and 2% of LACM 5833 and 5834 reads were classified as P. maniculatus. These numbers are comparable to other Neotominae off-target classifications and seem likely driven by conserved regions of the genome between related species, rather than cross-contamination or hybridization (Supplemental Data SD3).

Mitogenome phylogenetic analysis and molecular dating. In both BI and ML trees, LACM 5833 and LACM 5834 were recovered as sisters to each other and then sister to the P. truei reference mitogenome (USNM 603235), confirming the species identification as P. truei (Figure 3, Supplemental Data SD4-5). The reference genome is P. truei truei, which has previously been shown to belong to a separate clade from the western subspecies of P. truei (Hernández-Canchola et al. 2022). BEAST dated the split between the P. truei truei mitogenome and the LACM haplotypes to 512-670 kya, demonstrating significant distance between these two clades (Supplemental Data SD4).

LACM 6722 was recovered within the Peromyscus maniculatus species group (Clade P), sister to a clade of samples identified as P. maniculatus rufinus, P. maniculatus bairdii, and P. polionotus (1.0 posterior probability; 100 UF bootstrap). This position is consistent with previous placement of P. gambelii (Bradley et al. 2019; Greenbaum et al. 2019) and Channel Island mice specifically (Becker et al. 2025) within the Peromyscus maniculatus species group using cytochrome b. However, this position is not stable, since the relationship between P. maniculatus and P. polionotus is not resolved in the ML tree (92 UF bootstrap, Supplemental Data SD5), and in the BEAST tree, LACM 6722 is recovered as sister to the P. maniculatus mitogenomes only. This complex in general has been difficult to resolve and requires more loci to fully disentangle these relationships (Bradley et al. 2019; Castañeda-Rico et al. 2025). Nonetheless, BEAST dates the split between P. gambelii clementis and its sisters to 554-745 kya. This is notably a shorter time than the 1.8 mya split previously estimated with cytochrome b (Bradley et al. 2019).

Cytochrome b phylogeographic analysis. Phylogenetic analysis of P. truei cytochrome b haplotypes recovered the two major clades identified by Hernández-Canchola et al. (2022). The P. truei I clade includes the P. t. truei reference mitogenome from New Mexico and other specimens from Arizona, Colorado, New Mexico, Oklahoma, Texas, and Utah (Figure 4). The P. truei II clade comprises specimens from California, Idaho, Nevada, and Utah. While we did not examine these specimens to verify their subspecies assignment within each clade, the observed geographic structuring is consistent with previous findings (Hernández-Canchola et al. 2022). The P. truei I clade was not resolved in the ML tree (Supplemental Data SD6), but P. truei II was supported in both ML and BI trees.

LACM 5833 and LACM 5834 were recovered in the P. truei II clade, which also contained all other specimens collected in California. We did not recover fine phylogeographic structure within the major clades, with polytomies in both ML/BI trees, limiting our ability to identify a source population for the P. truei caught on SCL. LACM 5833 is sister to haplotypes associated with MVZ specimens caught in Berkeley, CA, as suggested by the BLAST search, but this relationship only had support in the BI tree. The phylogenetic position of LACM 5834 was not resolved within the P. truei II clade.

Morphological analysis. LACM 5833 and LACM 5834 are much larger than the other SCL Peromyscus we sequenced (Figure 2) and have diagnostic features associated with P. truei. Their total body sizes are over 200 mm in total length (201, 207), with high tail to body ratios (0.97, 1.0), large ears (22 mm each), and visibly denser fur at the tip of the tail (Figure 2A). While the teeth of LACM 5833 were too worn for analysis, LACM 5834 exhibits distinct mesolophs/accessory cusps on its molars (Figure 2B), consistent with P. truei dental characters (Ingles 1965). By contrast, the specimens genetically confirmed as P. gambelii clementis have a median total body size of 172 ± 5.2 mm, a tail to body ratio of 0.79 ± 0.04, an ear length of 17 ± 0.4 mm, no tail tufting, and no accessory cusps (Figure 2C). The median measurements for these values across the 277 SCL specimens are relatively consistent with these specimens, but with high standard deviations: 175 ± 12.2 mm total length, 0.80 ± 0.08 tail/body ratio, and 17 ± 2.0 mm ear length. Boxplots of this population demonstrate that abnormally large specimens are not limited to LACM 5833 and LACM 5834; there are a significant number of outliers with much larger total body lengths, tail ratios, and ear sizes (Figure 5A-C). In fact, this distribution is nearly bimodal, with few specimens of intermediate size, and the genetically identified P. truei specimens clustering with other outliers (Figure 5D). We did observe densely furred tail tips in some of these specimens, but this character is inconsistent and dependent on specimen preparation and preservation.

Across the Channel Islands, median measurements and distributions differed between island populations, consistent with prior morphological studies. While SCL specimens were either distributed solidly below or above the 0.9 tail/body length ratio used in differentiating P. truei and P. maniculatus (Ingles 1965), this value overlapped with the distribution of other islands. Santa Cruz Island contained the largest individuals, with total body size of genetically confirmed P. gambelii specimens comparable to the outliers on SCL. Some sporadic large specimens with long tails collected on Anacapa, San Nicolas, Santa Barbara, and Santa Catalina were also genetically confirmed as P. gambelii. By contrast, the median measurements of Santa Rosa Island deer mice were the smallest of any island, but also contained many large outlier specimens, none of which were sequenced. The measurements of these outliers were generally not as high as the SCL outliers and overlapped with distributions on other islands (Figure 5C). Nonetheless, the violin plot of Santa Rosa Island total body length was the only other island to show a more bimodal distribution like SCL (Figure 5D). Some outliers for total body length on SCL, Anacapa, San Miguel, and San Nicolas were pregnant when collected (Supplemental Data SD7A), but sexual dimorphism did not explain any general size distribution patterns (Supplemental Data SD8). Many, but not all, small-bodied outliers were recorded as immature, and three outliers for high total body length on SCL were also immature individuals (Supplemental Data SD7B).

Study skins housed at the Smithsonian National Museum of Natural History and collected before the CIBS contain few abnormally large specimens, except on Santa Cruz Island (Supplemental Data SD9), where large-bodied island mice have previously been documented. Of the six specimens from the Santa Barbara Museum of Natural History Museum collected during the CIBS, three were Onychomys specimens which had been incorrectly re-labeled as P. maniculatus (SBMNH:MAM:7621; SBMNH:MAM:9202-9203; Supplemental Data SD10). Of the Peromyscus specimens, two (SBMNH:MAM:7619, SBMNH:MAM:7620) were large-bodied (total length 198, 192 mm, respectively) with large ears (19 mm) and intermediate tail size (0.9, 0.88 tail/body ratio; Supplemental Data SD11). These measurements are consistent with descriptions of both P. maniculatus and P. truei (Ingles 1965). Neither specimen displayed dense fur at their tail tips of (Supplemental Data SD10), yet both were substantially larger than the median SCL island deer mouse.

Discussion

Mitochondrial genomes, low coverage nuclear data, and morphological characters of LACM 5833 and LACM 5834 confirm their identity as P. truei, not the endemic SCL deer mouse P. gambelii clementis. These specimens represent the first assembled mitogenomes from the western “P. truei Clade II” (P. cf. martirensis; Hernández-Canchola et al. 2022), which we estimate diverged from “P. truei Clade I” ~512-670 kya. Distinct mitogenome haplotypes indicate multiple individuals were introduced rather than a single pregnant female. Morphometric analysis of other CIBS specimens revealed additional individuals with ear and tail measurements consistent with P. truei, suggesting breeding P. truei populations on SCL in the 1930s. Although limited phylogeographic resolution at cytochrome b precludes pinpointing a source population of the detected invasion, the closest related haplotypes derived from mainland California P. truei. This pattern aligns with historical accounts of hay shipments introducing other rodents, including Microtus californicus and Reithrodontomys megalotis (von Bloeker Jr 1967), and underscores how human-mediated movements repeatedly reshaped island faunas.

Anthropogenic history of Channel Island rodents and San Clemente Island. The archaeological record shows that P. gambelii arrived on the Channel Islands after or concurrently with humans (Shirazi et al. 2018), suggesting that deer mice were inadvertently introduced some 13,000 years ago. Trade networks between the mainland and among the Island Chumash and Gabrielino/Tongva peoples living on the Channel Islands likely provided further opportunities for translocation (Johnson et al. 1983), which may also explain Holocene introductions of harvest mice (Reithrodontomys megalotis) on Santa Catalina and Santa Cruz Islands (Ashley 1989). Other mammals, such as the island fox and the Catalina ground squirrel, may have been intentionally translocated by Indigenous people (Collins 1991a; 1991b; Rick et al. 2009; 2019; 2024). However, in addition to these species which have lived on the Channel Islands for thousands of years, several rodents were repeatedly introduced during 19th-20th centuries along with other commensal and domestic animals: Rattus norvegicus on Catalina; Rattus rattus on Catalina, San Clemente, San Miguel, and Anacapa Island; Mus musculus on Catalina and San Clemente; Microtus californicus on Catalina and San Clemente, and Reithrodontomys megalotis on San Clemente (Rick 2013). While many islands experienced agricultural activity and mainland connectivity during this period, this especially rich rodent diversity on the SCL including adept island invaders like Rattus and Mus is striking.

While considered federal land since 1850, SCL was used primarily for private ranching operations until the US Navy assumed control in 1934. Today, Naval Base Coronado administers military operations on SCL, including extensive training facilities and the US Navy’s last remaining ship-to-shore live firing range. Increasing military development during the 1930s was a motivating factor for the CIBS team to begin expeditions on SCL as quickly as possible, according to their initial proposal for the survey (Laverty 2020). However, there had already been a long history of ecological deterioration on the island due to the impact of sheep and goat ranching. The San Clemente sheep company operated from at least 1879, grazing tens of thousands of sheep on the island. Feral goats, likely the most destructive of the animals introduced to SCL, were finally eradicated in 1991 by the US Navy, with over 29,000 total goats removed since 1972 (Keegan et al. 1994). CIBS botanist Meryl Dunkle wrote that, “The heavy grazing of sheep and goats in the past gave such a desolate appearance to the island that it seemed superficially of but little interest to the botanist” (Dunkle 1943). However, SCL is recognized today for harboring the highest number of endemic plants in the Channel Islands archipelago (Raven 1963), aided by Navy conservation and restoration efforts.

Today, proactive biosecurity measures are in place on the Channel Islands (e.g., Brenner et al. 2025), as stakeholders understand the potential for invasive species undermining meticulous conservation restoration efforts. Millions of dollars have been spent eradicating feral animals left behind from the ranching period, especially as these species have threatened charismatic endemics: feral goats on SCL degraded habitats for the loggerhead shrike (Lanius ludovicianus mearnsi) and island night lizard (Xantusia riversiana reticulata) (Scott and Morrison 1990; Keegan et al. 1994); feral ungulates on the northern Channel Islands indirectly led to hyperpredation on island foxes from golden eagles (Roemer et al. 2001; Coonan et al. 2014); feral cats on Santa Barbara and rats on Anacapa predated on the rare Scripps’s murrelet (Synthliboramphus scrippsi) (McChesney and Tershy 1998; Whitworth et al. 2013). Invasive species and habitat degradation on Anacapa and SCL have also impacted island deer mice, resulting in the State of California recognizing these two subspecies, P. g. anacapae and P. g. clementis, as critically imperiled (California Natural Diversity Database 2026). Monumental eradication efforts led by collaborations between National Park Services, the US Navy, and conservation organizations have led to the recovery and delisting of many Channel Islands species from the endangered species act, including the island fox and island night lizard (US Fish and Wildlife Service 2014; 2016), and recently on SCL, a unique Bell’s sparrow subspecies (Artemisiospiza belli clementeae) and four endemic plants (US Fish and Wildlife Service 2023).

Timing and pathways of introduction. Mammalogist Jack von Bloeker collected both native and introduced mammals during the CIBS, and attempted to infer their colonization histories in his comprehensive list of Channel Island mammals (von Bloeker Jr 1967). For example, he examined 28 M. californicus and 36 R. megalotis specimens from SCL, and determined that they were indistinguishable from geographically proximate mainland species (von Bloeker Jr 1967). He hypothesized that both of these species were inadvertently introduced by imported hay from San Diego County: “In April, 1939, I obtained information which supports this suggestion from a conversation with Mr. Theodore Murphy, United States Marshal, who was stationed on the island at that time and was the recipient of the hay as feed for his horse which he used in patrolling the island” (von Bloeker Jr 1967). These Microtus and Reithrodontomys were collected in geographic proximity to both confirmed P. gambelii and P. truei on southern SCL. In addition, six Onychomys specimens were also collected by von Bloeker in this area (LACM 6593- 6594, LACM 6596; SBMNH:MAM:7621, SBMNH:MAM:9202-9203), which are the only Onychomys individuals ever recorded on any of the Channel Islands (Knapp et al. 2021).

The co-localized collection of non-native Microtus, Reithrodontomys, and Onychomys on southern SCL indicates an active route for wild (not just commensal) rodent invasions from mainland California during the Channel Islands ranching era. Furthermore, von Bloeker’s correspondence with Theodore Murphy, who was ranching on SCL prior to serving as US Marshal, supports hay shipments as the introduction pathway for at least two of these species. If P. truei arrived contemporaneously, its introduction likely occurred after 1926 when Murphy began activity on SCL. This timeline is concordant with earlier Peromyscus specimens (n=37) collected on SCL between 1894-1897 by Mearns, Anthony, and Gaylord from the International Boundary Commission, which are phenotypically consistent with P. gambelii (Supplemental Data SD9). Because localities for these earlier specimens are unrecorded, and CIBS sampling was concentrated at the island’s southern end, our ability to infer the distribution or spread of P. truei remains limited.

Specimen identification and historical notes. Given von Bloeker’s extensive experience collecting and identifying rodents across California, it is notable that he did not classify any SCL specimens as P. truei. Specimens such as LACM 5833 and 5834 exhibit distinct tail tips, diagnostic dental characters, and body size significantly larger than both Mearns’s descriptions and von Bloeker’s average Peromyscus specimens. Unfortunately, von Bloeker’s complete field notes are unavailable, and because SCL mice had already been described as a subspecies, he did not formally analyze them as he did for Santa Rosa and Anacapa Islands. However, original tag labels suggest he may have recognized some of these large mice as atypical: while most were labeled “Peromyscus maniculatus clementis” or “Peromyscus m. clementis”, the original pen labels of LACM 5833, 5834, and other specimens were solely labeled “Peromyscus” (Figure 2). Later curatorial staff added species and subspecies epithets in pencil, but von Bloeker himself appears not to have resolved these identifications. Notes from a 1941 expedition to SCL further support this interpretation: “Jack von Bloeker secured … a considerable number of the San Clemente white-footed mouse, Pero-myscus maniculatus clementis. In addition, several examples of Peromyscus, Reithrodontomys and Microtus, not yet determined, were taken, and are being given further study” (Comstock 1946). Some pencil annotations even list the species epithet such as “boylii” or “californicus”, though no diagnostic characters for these species were observed in the specimens we examined. Nonetheless, we recommend molecular identification for these specimens, especially those with damaged teeth or without skulls, to definitively resolve this issue without further misidentifications.

Several unusually large Peromyscus specimens from Santa Rosa and Anacapa Islands share similar labeling patterns to those on San Clemente, with pen-labeled “Peromyscus” and pencil-marked “boylii”. On Anacapa, four such specimens (LACM 7394-7397) were genetically confirmed P. gambelii, along with other individuals with outlier total body lengths (Figure 5C). von Bloeker himself noted large body sizes and tail lengths in his description of the Anacapa subspecies (von Bloeker Jr 1941), and large body sizes and tail lengths on Santa Cruz Island are consistent with known subspecific traits (Nelson and Goldman 1931), with multiple outliers for total body length genetically confirmed as P. gambelii.

Santa Rosa Island presents a more ambiguous case. A significant series of large-bodied specimens collected in 1939 from Elderberry and Cherry Canyons were not genetically tested, and dental characters were inconclusive. von Bloeker described P. m. sanctaerosae as one of the smallest island mouse subspecies with uniquely short tails (von Bloeker Jr 1940), yet our morphological analysis identified several outliers exceeding 170 mm in total length. Comparing von Bloeker’s reported specimen counts to current LACM holdings suggests that these outliers were excluded from his subspecies description, possibly because he did not consider them P. m. sanctaerosae. Their collection near Vail Ranch, where calves were routinely shipped for nearly a century, provides a plausible introduction pathway via livestock transport. Without genetic data or field notes, the identity of these large Santa Rosa mice remains unresolved.

Although we detected evidence of non-native Peromyscus on San Clemente and possibly Santa Rosa Islands using standard measurements, morphological identification in this group is inherently challenging. Peromyscus is highly diverse, with unresolved species complexes and substantial phenotypic variation that complicates differentiation among sympatric lineages (Rich et al. 1996; Platt et al. 2015; Castañeda-Rico et al. 2025). Channel Island deer mice also exhibit mild island gigantism making them larger than mainland forms (Collins et al. 1979; Gill 1980). These size differences reduce the utility of identification keys based on mainland Peromyscus, as exemplified by the many genetically confirmed island mice with a tail to body ratio of exceeding 0.9 (Figure 5C). Morphological plasticity on the Channel Islands further complicates identification (Pergams and Ashley 1999), though we minimized this by restricting analyses to specimens collected during the same three-year survey period. While von Bloeker likely suspected that unusually large Peromyscus on San Clemente were atypical, similar suspicions may have applied to outlier Anacapa and Santa Cruz specimens.

Implications and further directions. LACM 5833, LACM 5834, and other specimens may not have been correctly identified as P. truei because this species is not typically associated with island invasions. In fact, insular forms of P. truei, native or introduced, have never been reported. While some subspecies specialize in piñon-juniper habitats, California P. truei occupy a diversity of habitats, including chaparral on canyon slopes (Hoffmeister 1981), suggesting suitable habitat exists on the Channel Islands, even on comparatively depauperate SCL. Limited post-CIBS collecting makes it unclear whether P. truei populations persisted past the 1940s. Although P. truei is absent from published records and platforms like iNaturalist (Knapp et al. 2021), it is likely that Channel Island rodent identifications are rarely scrutinized beyond the generic level with only one known local Peromyscus species. Determining whether P. truei remains extant on SCL is critical, as introduced rodents may thrive in degraded habitats and threaten the critically imperiled endemic P. gambelii clementis. Monitoring is essential as restoration progresses to assess shifts in rodent communities. While eradication of black rats on Anacapa succeeded without negatively affecting native island deer mice populations (Howald et al. 2010; Ozer et al. 2011), similar efforts for P. truei on SCL would be impractical due to the island’s large size and P. truei’s morphological similarity to native island deer mice.

Over 80 years after the CIBS, museum genomics has resolved long standing misidentifications within Peromyscus, confirming that large-bodied Anacapa and Santa Cruz specimens belong to endemic subspecies, while revealing a cryptic invasion of P. truei on SCL. Cytochrome b comparisons link these specimens to mainland California populations, consistent with the hypothesis that hay bales imported from San Diego during the 1930s may have inadvertently translocated multiple species of wild rodents (von Bloeker Jr 1967). Morphological evidence suggests P. truei may have formed substantial populations on SCL, and perhaps on Santa Rosa Island. We recommend molecular screening and comprehensive morphological analysis in collections and targeted field surveys to test identifications and better understand the scope of these introductions. Accurate identification is essential for conservation management on the Channel Islands, and for collections-based research in this taxonomically complex group.

Acknowledgements

This manuscript is dedicated to Dr. Livia León-Paniagua, whose research, mentorship, and leadership have profoundly shaped the field of mammalogy and created a lasting legacy in the study of Mexican mammals. We are deeply grateful for her unwavering commitment to training new generations of mammalogists and for the collaborative spirit that has distinguished her work with students and colleagues. Thanks also to LACM for a Student Collections Study Award which enabled initial sampling of museum specimens for this study. Additional thanks to Shannen Robson (LACM), Melissa Hawkins and Darrin Lunde (USNM), and Ally Coconis and Krista Fahy (SBMNH) for access to the collections and specimens examined. This work was supported by CCG, Smithsonian National Zoo and Conservation Biology Institute federal funding, the Smithsonian-Mason School of Conservation, George Mason University, an American Genetics Association EECG Award, the American Society of Mammalogists Grants-in-Aid and Horner Awards, a Society of Systematic Biologists Graduate Student Research Award, an American Wildlife Conservation Foundation Grant, and a Cosmos Scholar Grant. Digitization of specimen traits from study skin tags was funded by the National Science Foundation (DBI 2228388 to KCB). MAB was also supported by an SMSC Fellowship through the Jacquemin Family Foundation and by the George Mason University Graduate Division. This project is supported in part by federal funding from the Latino Initiatives Pool, administered by the National Museum of the American Latino. This funding supports the UCSB-Smithsonian Scholars Program and SLO’s internship. Thanks to Mario Castellanos and Melani Castellanos for support of UCSB-Smithsonian Scholars. Additional thanks to Nancy McInerney and CCG lab members for support, including feedback on this project. We are grateful to Jesús A. Fernández and an anonymous reviewer for providing valuable input and improving the quality of this manuscript.

Declaration of Artificial Intelligence use

No generative AI tools were used in the preparation of this manuscript.

Author contributions

Madeleine A. Becker: Conceptualization, Formal Analysis, Funding Acquisition, Investigation, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing; Kayce C. Bell: Data Curation, Investigation, Writing – Review & Editing; Selena Lopez-Ortiz: Formal Analysis, Visualization, Writing – Review & Editing; Jesús E. Maldonado: Conceptualization, Funding Acquisition, Resources, Supervision, Writing – Review & Editing; Cody W. Edwards: Funding Acquisition, Supervision, Writing – Review & Editing; Susette Castañeda-Rico: Conceptualization, Formal Analysis, Supervision, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing.

Supplementary data

SD1. A) Measurements from Natural History Museum of Los Angeles County (LACM) Channel Island Peromyscus specimens collected during the Channel Islands Biological Survey (1939-1941), which were genetically sampled and morphologically examined. All metadata are derived from original study skin tags, except for tail to body size ratio, which was calculated from these measurements. B) Measurements from Natural History Museum of Los Angeles County (LACM) Channel Island Peromyscus specimens collected during the Channel Islands Biological Survey (1939-1941), which were morphologically examined but not genetically sampled. All metadata are derived from original study skin tags, except for tail to body size ratio, which was calculated from these measurements. C) GenBank and Sequence Read Archive accession numbers for annotated mitochondrial genomes, novel cytochrome b haplotypes, and shotgun genomic data generated.

SD2. Shotgun sequencing and mito-chondrial genome assembly statistics for San Clemente Island P. truei specimens (LACM 5833, LACM 5834), San Clemente Island mouse (P. gambelii clementis; LACM 6722), and reference P. truei truei (USNM 603235).

SD3. Kronagrams displaying classi-fication results of low coverage whole genome data, restricted to reads classified as Peromyscus at the genus level: A) LACM 5833 (P. truei), B) LACM 5834 (P. truei), and C) LACM 6722 (P. gambelii clementis).

SD4. Time-scaled mitogenome tree inferred by BEAST, showing the positions of LACM 5833 and LACM 5834 (P. truei) and LACM 6722 (P. gambelii clementis), subset to the tribe Peromyscini and its sister genus (Onychomys). Nodes are labeled with median estimated divergence time in millions of years, with green bars displaying the 95% confidence interval.

SD5. Maximum Likelihood mito-genome consensus tree showing positions of LACM 5833 and LACM 5834 (P. truei), and LACM 6722 (P. gambelii clementis), subset to the tribe Peromyscini and its sister genus (Onychomys). The Peromyscus truei species group is highlighted in red, and the Peromyscus maniculatus species group is highlighted in blue. Clade names and sequences are replicated from Castañeda-Rico et al. 2025. Nodes with black dots indicate high Ultrafast bootstrap support >95.

SD6. Maximum Likelihood cytochrome b consensus tree showing positions of LACM 5833 and LACM 5834 within Peromyscus truei. Clades are labeled in accordance with Hernández-Canchola et al. (2022). The P. truei II clade includes all haplotypes sequenced from California specimens. This phylogenetic tree has been subset to P. truei and its sister (P. gratus) for relevance. Nodes with black dots indicate high Ultrafast bootstrap support >95.

SD7. Distribution of study skin measurements from Channel Islands Biological Survey specimens (n=701) prepared by Jack von Bloeker. Boxplots of total body size by island (mm) with A) pregnant females foregrounded according to reproductive status as originally recorded on tags by von Bloeker, B) immature individuals foregrounded according to age class as originally recorded on tags by von Bloeker.

SD8. Distribution of study skin measurements from Channel Islands Biological Survey specimens (n=701) prepared by Jack von Bloeker. Boxplots of total body size (mm) by island and by sex.

SD9. Measurements from National Museum of Natural History (USNM) Channel Island Peromyscus specimens collected before the Channel Islands Biological Survey. All metadata are derived from original study skin tags, except for tail to body size ratio, which was calculated from these measurements. Many specimen tags did not include complete metadata, resulting in missing data.

SD10. Santa Barbara Museum of Natural History (SBMNH) Peromyscus (SBMNH:MAM:7619-7620, SBMNH:MAM:7622) and Onychomys (SBMNH:MAM:7621, SBMNH:MAM:9202-9203;) specimens originally collected on San Clemente Island during the Channel Islands Biological Survey. Onychomys specimens had been erroneously re-labeled as P. maniculatus in collections.

SD11. Measurements of Santa Barbara Museum of Natural History (SBMNH) Peromyscus and Onychomys specimens originally collected on San Clemente Island during the Channel Islands Biological Survey. Onychomys specimens had previously been re-labeled as P. maniculatus in collections. All metadata are derived from original study skin tags, except for tail to body size ratio, which was calculated from these measurements.

Literature cited

Ashley MV. 1989. Absence of differentiation in mitochondrial DNA of island and mainland harvest mice, Reithrodontomys megalotis. Journal of Mammalogy 70:383–386. https://doi.org/10.2307/1381523

Ashley MV, and Wills C. 1989. Mitochondrial-DNA and allozyme divergence patterns are correlated among island deer mice. Evolution 43:646–650.

Becker MA, Carrasco C, Kautt AF, Castañeda-Rico S, Orrock JL, Maldonado JE, et al. 2025. Phylogenetics of Channel Island deer mice based on the cytochrome b gene sheds light on multiple colonization events and supports current taxonomy. Western North American Naturalist 85:293–309. https://doi.org/10.3398/064.085.0215

Becker MA. In preparation. Evolutionary Genomics of Channel Island Peromyscus. [PhD thesis]. [Virginia (EEUU)]: George Mason University.

Benham PM, and Bowie RCK. 2022. Natural history collections as a resource for conservation genomics: Understanding the past to preserve the future. Journal of Heredity 114:367–384. https://doi.org/10.1093/jhered/esac066

Bouckaert R, Vaughan TG, Barido-Sottani J, Duchêne S, Fourment M, Gavryushkina A, et al. 2019. BEAST 2.5: an advanced software platform for Bayesian evolutionary analysis. PLoS Computational Biology 15(e1006650). https://doi.org/10.1371/journal.pcbi.1006650

Bradley RD, Francis JQ, Platt II RN, Soniat TJ, Alvarez D, and Lindsey LL. 2019. Mitochondrial DNA sequence data indicate evidence for multiple species within Peromyscus maniculatus. Special Publications, Museum of Texas Tech University 70:1–59.

Brenner LJ, Rindlaub N, Matos J, Meyler S, Pollock S, Schuetzenmeister F, et al. 2025. Real-time island biosecurity surveillance: Evaluating a wireless camera network for AI-assisted early detection of invasive mammals on Santa Cruz Island, CA. Western North American Naturalist 85:365–375. https://doi.org/10.3398/064.085.0220

California Natural Diversity Database [CNDDB]. 2026. Special Animals List. Sacramento (CA, EEUU): California Department of Fish and Wildlife; April 2026. [Accessed May 11th, 2026]. https://nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=109406

Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, et al. 2009. BLAST+: Architecture and applications. BMC Bioinformatics 10:421. https://doi.org/10.1186/1471-2105-10-421

Castañeda-Rico S, Maldonado JE, Hawkins MTR, and Edwards CW. 2025. Unveiling hidden diversity: Phylogenomics of neotomine rodents and taxonomic implications for the genus Peromyscus. Molecular Phylogenetics and Evolution 203:108233. https://doi.org/10.1016/j.ympev.2024.108233

Collins PW, Storrer J, and Rindlaub K. 1979. XI. Vertebrate zoology: biology of the deer mouse, A natural resources study of the Channel Islands National Park, California. In: Power DM, editor. Final Technical Report to the Denver Service Center, National Park Service, Denver, CO (EEUU); p. 11.1–11.74.

Collins PW. 1991a. Interaction between island foxes (Urocyon littoralis) and Indians on islands off the coast of southern California: I. Morphologic and archaeological evidence of human assisted dispersal. Journal of Ethnobiology 11:51–81.

Collins PW. 1991b. Interaction between island foxes (Urocyon littoralis) and Native Americans on islands off the coast of southern California: II. Ethnographic, archaeological, and historical evidence. Journal of Ethnobiology 11:205–229.

Comstock JA. 1946. Contributions from the Los Angeles museum—Channel Islands Biological Survey--33. Brief notes on the expeditions conducted between March 16, 1940 and December 14, 1941. Bulletin, Southern California Academy of Sciences 45:94–107.

Coonan TJ, Bakker V, Hudgens B, Boser CL, Garcelon DK, and Morrison SA. 2014. On the fast track to recovery: Island foxes on the northern Channel Islands. Monographs of the Western North American Naturalist 7:373–381. https://doi.org/10.3398/042.007.0128

Dunkle MB. 1943. Contributions from the Los Angeles Museum-Channel Islands Biological Survey No. 28: Three new plants from San Clemente Island. Bulletin of the Southern California Academy of Sciences 42:31–33.

Elliot DG. 1903. Descriptions of apparently new species and subspecies of mammals from California, Oregon, the Kenai Peninsula, Alaska, and Lower California, Mexico. Field Columbian Museum Publication 74:141–149.

Erlandson JM, Rick TC, Braje TJ, Casperson M, Culleton B, Fulfrost B, et al. 2011. Paleoindian seafaring, maritime technologies, and coastal foraging on California’s Channel Islands. Science 331:1181–1185. https://doi.org/10.1126/science.1201477

Gill AE. 1980. Evolutionary genetics of California islands Peromyscus. In: Power DM, editor. The California Islands: Proceedings of a multidisciplinary symposium. Santa Barbara Museum of Natural History, Santa Barbara, CA (EEUU); p. 719–743.

Goedknegt MA, Thieltges DW, Van Der Meer J, Wegner KM, and Luttikhuizen PC. 2018. Cryptic invasion of a parasitic copepod: Compromised identification when morphologically similar invaders co-occur in invaded ecosystems. PLoS ONE 13:e0193354. https://doi.org/10.1371/journal.pone.0193354

Greenbaum IF, Chirhart SE, Walker ML, and Honeycutt RL. 2017. Molecular phylogenetics of western deer mice (Peromyscus): Taxonomic and biogeographic implications. The Southwestern Naturalist 62:129–137. https://doi.org/10.1894/0038-4909-62.2.129

Greenbaum IF, Honeycutt R, and Chirhart S. 2019. Taxonomy and phylogenetics of the Peromyscus maniculatus species group. Special Publications, Museum of Texas Tech University 71:559–575.

Hernández-Canchola G, León-Paniagua L, and Esselstyn JA. 2022. Mitochondrial DNA and other lines of evidence clarify species diversity in the Peromyscus truei species group (Cricetidae: Neotominae). Mammalia 86:380–392. https://doi.org/10.1515/mammalia-2021-0146

Hoffmeister DF. 1981. Peromyscus truei. Mammalian Species 161:1–5.

Howald GR, Donlan CJ, Faulkner KR, Ortega S, Gellerman H, Croll DA, et al. 2010. Eradication of black rats Rattus rattus from Anacapa Island. Oryx 44:30–40. https://doi.org/10.1017/S003060530999024X

Ingles LG. 1965. Mammals of the pacific states: California, Oregon, and Washington. Redwood City (EEUU): Stanford University Press.

Johnson DL. 1983. The California continental borderland: Landbridges, watergaps and biotic dispersals. In: Master PM, and Flemming NC, editors. Quaternary Coastlines and Marine Archaeology: Towards the Prehistory of Land Bridges and Continental Shelves. Academic Press, London (UK) and New York, NY (EEUU); p. 481–527.

Kapp JD, Green RE, and Shapiro B. 2021. A fast and efficient single-stranded genomic library preparation method optimized for ancient DNA. Journal of Heredity 112:241–249. https://doi.org/10.1093/jhered/esab012

Katoh K, and Standley DM. 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution 30:772-780. https:// doi.org/10.1093/molbev/mst010

Keegan DR, Coblentz BE, and Winchell CS. 1994. Feral goat eradication on San Clemente Island, California. Wildlife Society Bulletin 22:56–61.

Kim AS, Kreiner JM, Hernández F, Bock DG, Hodgins KA, and Rieseberg LH. 2023. Temporal collections to study invasion biology. Molecular Ecology 32:6729–6742. https://doi.org/10.1111/mec.17176

Knapp DA, Collins PW, and Etter K. 2021. Channel Islands mammals: A gap analysis of specimen and observation data. Unpublished final report prepared for The Nature Conservancy by the Santa Barbara Botanic Garden, Santa Barbara, California. Santa Barbara (CA, EEUU): Santa Barbara Botanic Garden; October, 2021. [Accessed May 11th, 2026]. https://sbbotanicgarden.org/wp-content/uploads/2022/06/MAMMALS-gap-analysis-final.pdf

Lanfear R, Frandsen PB, Wright AM, Senfeld T, and Calcott B. 2017. PartitionFinder 2: New methods for selecting partitioned models of evolution for molecular and morphological phylogenetic analyses. Molecular Biology and Evolution 34:772–773. https://doi.org/10.1093/molbev/msw260

Laverty CH. 2020. North America’s Galapagos: The Historic Channel Islands Biological Survey. Salt Lake City (EEUU): University of Utah Press.

Light JE, Siciliano-Martina L, Dohlanik E, Vielleux G, Hafner D, Lawing AM, et al. 2021. Morphological differentiation of Peromyscus leucopus and P. maniculatus in East Texas. Therya 12:369–387. https://doi.org/10.12933/therya-21-1116

Lu J, Rincon N, Wood DE, Breitwieser FP, Pockrandt C, Langmead B, et al. 2022. Metagenome analysis using the Kraken software suite. Nature Protocols 17:2815–2839. https://doi.org/10.1038/s41596-022-00738-y

Martin M. 2011. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.Journal 17:10–12. https://doi.org/10.14806/ej.17.1.200

McChesney GJ, and Tershy BR. 1998. History and status of introduced mammals and impacts to breeding seabirds on the California Channel and northwestern Baja California islands. Colonial Waterbirds 21:335–347. https://doi.org/10.2307/1521646

McDonough MM, Parker LD, Rotzel McInerney N, Campana MG, and Maldonado JE. 2018. Performance of commonly requested destructive museum samples for mammalian genomic studies. Journal of Mammalogy 99:789–802. https://doi.org/10.1093/jmammal/gyy080

Mearns EA. 1897. Descriptions of six new mammals from North America. Proceedings of the United States National Museum 19:719–724.

Mearns EA. 1907. Mammals of the Mexican boundary of the United States. United States National Museum Bulletin 56:1–530.

Meineke EK, Davies TJ, Daru BH, and Davis CC. 2019. Biological collections for understanding biodiversity in the Anthropocene. Philosophical Transactions of the Royal Society B: Biological Sciences 374:20170386. https://doi.org/10.1098/rstb.2017.0386

Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, et al. 2020. IQ-TREE 2: New models and efficient methods for phylogenetic inference in the genomic era. Molecular Biology and Evolution 37:1530–1534. https://doi.org/10.1093/molbev/msaa015

Nelson EW, and Goldman EA. 1931. Six new white-footed mice (Peromyscus maniculatus group), from islands off the Pacific Coast. Journal of the Washington Academy of Sciences 21:530–535.

O’Hanlon SJ, Rieux A, Farrer RA, Rosa GM, Waldman B, Bataille A, et al. 2018. Recent Asian origin of chytrid fungi causing global amphibian declines. Science 360:621–627. https://doi.org/10.1126/science.aar1965

Ondov BD, Bergman NH, and Phillippy AM. 2011. Interactive metagenomic visualization in a Web browser. BMC Bioinformatics 12:385. https://doi.org/10.1186/1471-2105-12-385

Ozer F, Gellerman H, and Ashley MV. 2011. Genetic impacts of Anacapa deer mice reintroductions following rat eradication: Anacapa island mouse reintroduction. Molecular Ecology 20:3525–3539. https://doi.org/10.1111/j.1365-294X.2011.05165.x

Pergams ORW, and Ashley MV. 1999. Rapid morphological change in Channel Island deer mice. Evolution 53:1573–1581. https://doi.org/10.1111/j.1558-5646.1999.tb05420.x

Pergams ORW, and Ashley MV. 2002. California Island deer mice: Genetics, morphometrics, and evolution. In: Browne DR, Mitchell KL, and Chaney HW, editors. Proceedings of the Fifth California Islands Symposium. Santa Barbara, CA (EEUU): Santa Barbara Museum of Natural History; p. 278–288.

Pergams ORW, Byrn D, Lee KLY, and Jackson R. 2015. Rapid morphological change in black rats (Rattus rattus) after an island introduction. PeerJ 3:e812. https://doi.org/10.7717/peerj.812

Platt RN, Amman BR, Keith MS, Thompson CW, and Bradley RD. 2015. What Is Peromyscus? Evidence from nuclear and mitochondrial DNA sequences suggests the need for a new classification. Journal of Mammalogy 96:708–719. https://doi.org/10.1093/jmammal/gyv067

Provan J, Booth D, Todd NP, Beatty GE, and Maggs CA. 2008. Tracking biological invasions in space and time: Elucidating the invasive history of the green alga Codium fragile using old DNA. Diversity and Distributions 14:343–354. https://doi.org/10.1111/j.1472-4642.2007.00420.x

R Core Team. 2021. R: A language and environment for statistical computing. R Foundation for Statistical Computing. V. ٢.١٤.٢. Vienna (Austria). https://www.R-project.org/.

Raven PH. 1963. A flora of San Clemente Island, California. Aliso: A Journal of Systematic and Floristic Botany 5:289–347.

Rambaut A, Drummond AJ, Xie D, Baele G, and Suchard MA. 2018. Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Systematic Biology 67:901–904. https://doi.org/10.1093/sysbio/syy032

Raxworthy CJ, and Smith BT. 2021. Mining museums for historical DNA: Advances and challenges in museomics. Trends in Ecology and Evolution 36:1049–1060. https://doi.org/10.1016/j.tree.2021.07.009

Rich SM, Kilpatrick CW, Shippee JL, and Crowell KL. 1996. Morphological differentiation and identification of Peromyscus leucopus and P. maniculatus in northeastern North America. Journal of Mammalogy 77:985–991. https://doi.org/10.2307/1382779

Rick TC. 2013. Hunter-gatherers, endemic island mammals, and the historical ecology of California’s Channel Islands. In: Thompson VD, and Waggoner JC, editors. The Archaeology and Historical Ecology of Small Scale Economies. Gainseville (EEUU): University Press of Florida; p. 41–64. https://doi.org/10.5744/florida/9780813042428.003.0003

Rick TC, Erlandson JM, Vellanoweth RL, Braje RJ, Collins PW, Guthrie DA, et al. 2009. Origins and antiquity of the island fox (Urocyon littoralis) on California’s Channel Islands. Quaternary Research 71:93–98. https://doi.org/10.1016/j.yqres.2008.12.003

Rick TC, Hofman CA, and Reeder-Myers LA. 2019. Why Translocate? Evaluating the evidence and reasons for ancient human introductions of wildlife to California’s Islands. In: Gill KM, Fauvelle M, Erlandson JM, editors. An Archaeology of Abundance: Reevaluating the Marginality of California’s Islands. Gainseville (EEUU): University Press of Florida; p. 248–272. https://doi.org/10.5744/florida/9780813056166.003.0009

Rick TC, Radde HD, Teeter WG, Elliott Smith EA, Alvitre CM, Dagtas ND, et al. 2024. Enhancing biodiversity: Historical ecology and biogeography of the Santa Catalina Island ground squirrel, Otospermophilus beecheyi nesioticus. Royal Society Open Science 11:240726. https://doi.org/10.1098/rsos.240726

Roemer GW, Coonan TJ, Garcelon DK, Bascompte J, and Laughrin L. 2001. Feral pigs facilitate hyperpredation by golden eagles and indirectly cause the decline of the island fox. Animal Conservation 4:307–318. https://doi.org/10.1017/S1367943001001366

Ronquist F, and Huelsenbeck JP. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574. https://doi.org/10.1093/bioinformatics/btg180

Schmieder R, and Edwards R. 2011. Quality control and preprocessing of metagenomic datasets. Bioinformatics 27:863–864. https://doi.org/10.1093/bioinformatics/btr026

Scott TA, and Morrison ML. 1990. Natural history and management of the San Clemente Loggerhead Shrike. Proceedings of the Western Foundation of Vertebrate Zoology 4:23–57.

Shirazi S, Rick TC, Erlandson JM, and Hofman CA. 2018. A tale of two mice: A trans-Holocene record of Peromyscus nesodytes and Peromyscus maniculatus at Daisy Cave, San Miguel Island, California. The Holocene 28:827–833. https://doi.org/10.1177/0959683617744266

Suarez AV, and Tsutsui ND. 2004. The value of museum collections for research and society. BioScience 54:66–74. https://doi.org/10.1641/0006-3568(2004)054[0066:TVOMCF]2.0.CO;2

US Fish and Wildlife Service [USFWS]. 2014. Final rule removing the island night lizard from the federal list of endangered and threatened wildlife. United States Federal Register 79:18190–18210. https://www.fws.gov/sites/default/files/federal_register_document/2014-06576.pdf

US Fish and Wildlife Service [USFWS]. 2016. Final rule: Removing the San Miguel island fox, Santa Rosa island fox, and Santa Cruz island fox from the federal list of endangered and threatened wildlife, and reclassifying the Santa Catalina island fox from endangered to threatened. United States Federal Register 81:53315–53333. https://www.fws.gov/sites/default/files/federal_register_document/2016-18778.pdf

US Fish and Wildlife Service [USFWS]. 2023. Endangered and threatened wildlife and plants; removing five species that occur on San Clemente Island from the federal lists of endangered and threatened wildlife and plants. United States Federal Register 88:4761–4792. https://www.fws.gov/sites/default/files/federal_register_document/2023-01400.pdf

von Bloeker Jr JC. 1940. Contributions from the Los Angeles Museum-Channel Islands Biological Survey No. 11: A new race of white-footed mouse from Santa Rosa Island, California. Bulletin of the Southern California Academy of Sciences 39:172–174.

von Bloeker Jr JC. 1941. Contributions from the Los Angeles Museum-Channel Islands Biological Survey No. 22: A new subspecies of white-footed mouse from the Anacapa Islands, California. Bulletin of the Southern California Academy of Sciences 40:161–162.

von Bloeker Jr JC. 1967. Land mammals of the southern California islands. In Philbrick RN, editor. Proceedings of the Symposium on the Biology of the California Islands. Santa Barbara, CA (EEUU): Santa Barbara Botanic Garden; p. 245–264.

Whitworth DL, Carter HR, and Gress F. 2013. Recovery of a threatened seabird after eradication of an introduced predator: Eight years of progress for Scripps’s murrelet at Anacapa Island, California. Biological Conservation 162:52–59. https://doi.org/10.1016/j.biocon.2013.03.026

Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, et al. 2019. Welcome to the Tidyverse. Journal of Open Source Software 4:1686. https://doi.org/10.21105/joss.01686

Wood DE, and Salzberg SL. 2014. Kraken: Ultrafast metagenomic sequence classification using exact alignments. Genome Biology 15:R46. https://doi.org/10.1186/gb-2014-15-3-r46

Yu G, Smith DK, Zhu H, Guan Y, and Lam TT. 2017. ggtree: An r package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods in Ecology and Evolution 8:28–36. https://doi.org/10.1111/2041-210X.12628

Associated editors: Giovani Hernández Canchola and Pablo Colunga Salas

Submitted: December 19, 2025; Reviewed: March 22, 2026

Accepted: May 8, 2026; Published online: May 29, 2026

THERYA, 2026, Vol. 17(2):277-294

DOI: 10.12933/therya.2026.6265 ISSN 2007-3364

Figure 1. Collecting localities of 287 Peromyscus specimens from San Clemente Island, USA, during the Channel Islands Biological Survey (1939-1941). Dot size indicates the number of specimens collected at each site. Colors denote whether sequenced specimens were genetically identified as endemic P. gambelii clementis or introduced P. truei. Inset: Map of the Channel Islands Archipelago relative to mainland southern California.

Figure 2. A) Peromyscus specimens from San Clemente Island, USA, collected during the Channel Islands Biological Survey (1939-1941) and housed at the Natural History Museum of Los Angeles County. Specimens LACM 5833 and LACM 5834 (second and third from the top, respectively) were re-identified as P. truei with genomic data. These specimens show morphological traits distinguishing them from the endemic P. gambelii, including longer tails with more densely furred tips, larger total body sizes, and larger ears. Original identifications from collector and CIBS mammalogist Jack von Bloeker appear in pen, while the specific and subspecific epithets in pencil were likely erroneously added later to LACM 5833 and LACM 5834. B) Upper molars of LACM 5834 exhibit diagnostic accessory cusps associated with P. truei (white arrows). C) Upper molars of LACM 6722 do not include these accessory cusps (absence indicated by white arrows), consistent with P. gambelii identification.

Figure 3. Bayesian mitogenome consensus tree showing positions of LACM 5833 and LACM 5834 (P. truei) and LACM 6722 (P. gambelii clementis), subset to the tribe Peromyscini and its sister genus (Onychomys). The Peromyscus truei species group is highlighted in red, and the Peromyscus maniculatus species group is highlighted in blue. Clade names and sequences are replicated from Castañeda-Rico et al. 2025. Nodes with black dots indicate high support with posterior probabilities >0.95.

Figure 4. Bayesian cytochrome b consensus tree showing positions of LACM 5833 and LACM 5834 within Peromyscus truei. Clades are labeled in accordance with Hernández-Canchola et al. (2022). The P. truei II clade includes all haplotypes sequenced from California specimens. This phylogenetic tree has been subset to P. truei and its sister (P. gratus) for relevance. Nodes with black dots indicate high support with posterior probabilities >0.95.

Figure 5. Distribution of study skin measurements recorded from Channel Islands Biological Survey specimens (n=701) prepared by Jack von Bloeker, including 75 genetically identified individuals, endemic P. gambelii (blue) and introduced P. truei (red), across each of the Channel Islands. A) Boxplot of ear length (mm). B) Boxplot of tail length to body length ratio; dashed line at 0.9 indicates threshold distinguishing P. truei from P. gambelii (Ingles 1965). C) Boxplot of total body length including tail (mm). D) Violin plot of total body length including tail (mm).