We constructed phylogenies of the 28S nuclear locus (figure 1) and the cox1 mitochondrial locus (figure 2) separately in order to include as many taxa as possible. These are the most frequently genotyped loci in Porifera, and these phylogenies are the most comprehensive to date for the order Suberitida. A recently published super matrix analysis 4 with more genetic data but many fewer species 4 is highly congruent with our phylogenies (Pankey et al. 2022) .

The type species Suberites domuncula (Olivi 1792) forms a highly supported clade containing a world-wide sample of other Suberites (clade 1, blue in figures 1 & 2). This clade includes the species known to be associated with hermit crabs, including Suberites latus, the hermit crab-associated representative from the North Pacific. There is a deep divergence between this Suberites 'sensu stricto' clade and clade 2, shown in red, which contains the remaining Suberites and the genus Rhizaxinella. Rhizaxinella are differentiated from Suberites by their stalked morphology: data from these two loci are insufficient to resolve whether stalks evolved only once in this clade. It is clear, however, that the genus Suberites contains at least two welldifferentiated clades, and it seems unlikely that spicule-based morphological characters will be able to resolve membership in these clades.

The North Pacific Homaxinella amphispicula (de Laubenfels 1961) is very morphologically similar to the North Atlantic H. subdola (Bowerbank 1866) , and DNA supports a close relationship at the 28S locus. The placement of the genera Homaxinella and Aaptos are inconsistent between loci, and will require additional data to resolve.

The remaining California suberitid species appear more closely related to the Halichondriidae than to Suberites. Pseudosuberites latke sp. nov. forms a clade with Pseudosuberites from South Africa, Antarctica, and Korea at the cox1 locus (clade 3, light blue). The type species of the genus, P. hyalinus Ridley & Dendy 1887 , is found in a different location in the cox1 tree, but this result should be considered tentative. The sequence is identical to a sample from the same collection that is identified as a Plicatellopsis, which may indicate some cross-contamination of DNA in that study. Several other samples identified as Pseudosuberites are found in our 28S tree, however, and also do not form a clade with the new species. This genus is therefore likely polyphyletic and in need of revision.

Thinly encrusting Suberitidae are placed in either Terpios (which have a gelatinous consistency, lobate tylostyles & low spicule density) or Protosuberites (with upright tylostyles) (Morrow & Cárdenas 2015; van Soest 2002) . The majority of Protosuberites fall within a single clade, shown in orange (clade 4, orange). The type species for the genus has not been sequenced, however, and the genus is polyphyletic due to Protosuberites denhartogi (van Soest & De Kluijver 2003) and Protosuberites ectyoninus (Topsent 1900) falling in other parts of the tree. The difficulty of identification in these morphologically simple sponges is highlighted by the observations that P. denhartogi appears to be a species complex based on 28S data, and samples identified as Terpios gelatinosa (Bowerbank 1866) fall in two different clades.

Suberites latus Lambe, 1893

Ficulina suberea var. lata (de Laubenfels 1932) Choanites suberea var. lata (de Laubenfels 1961) Suberites ficus (Bakus & Green 1987) Material examined. UF4128/BHAK-02603, UF4127/BHAK-02601, UF47673/BHAK-02604, UF4129/BHAK-02605 & UF4131/BHAK-02611, all collected at Rattenbury Pinnacle, British Columbia, Canada, (51.691, -120.080), 21 m, 8/3/17; UF4143/BHAK-03126, Big Spring Creek, British Columbia, Canada, (51.6501, -128.069), 2 m, 8/6/17; BHAK-10221 & BHAK-10390, both collected at Wedgeborough Point, British Columbia, Canada, (51.647, -127.956), 15 m, 5/27/19.

Morphology. Massive, amorphous sponges, nearly always containing a hermit crab. Color in life variable from vivid orange to reddish brown. Firm and smooth to the touch.

Skeleton. Ectosomal skeleton of tylostyles in upright bouquets. Tylostyles in choanosome in confusion with no apparent pattern.

Spicules. Tylostyles in two size classes, with spined centrotylote strongyles as microscleres. Short tylostyles are restricted to ectosome, where some long tylostyles can also be found. Tylostyle size classes divided at 200 ¿m following Austin et al. (2014) . Data are from two individuals from the Rattenbury Pinnacle collection, which were loaned in a common container. Long tylostyles: those we examined were nearly always smoothly curved or with a slight bend near the center (figure 3C), but a previous report found them to sometimes be straight or sinuous (Austin et al. 2014) . Most taper abruptly to a sharp point, but rounded points or strongylote endings present on a substantial minority. Heads usually terminal, round, and symmetrical, but occasionally subterminal. Shafts are often slightly fusiform, with thickest section in center (figure 3C), but some have consistent thickness throughout. 21633053409 x 337311 ¿m (n=95) Short tylostyles: shaped the same as long tylostyles. 10331583198 x 23739 ¿m (n=67) All tylostyles considered together: 10332453409 x 237311 ¿m (n=162), modes at approximately 170 and 310 ¿m.

Centrotylote microstrongyles: variable abundance from rare to common, and apparently sometimes absent. Covered in very small spines which are difficult to see in a light microscope (figure 3C). Tylote ball is generally near the center but sometimes at one end. Shaft sometimes straight but often curved or bent at the center (figure 3F). Austin et al. (2014) report that these spicules are choanosomal. Consistent with this report, they were much more common in the choanosome in one sample we examined, and very rare in both choanosome and ectosome in the other sample we examined. 14324333 x 13436 ¿m (n=84 length, n=34 width). Distribution and habitat. This species is known from a wide depth range (63183 m) from the Aleutian Islands, Alaska to British Columbia (Austin et al. 2014) . It is nearly always found in an association with a hermit crab, either on a gastropod shell or in the shape of a gastropod shell that has presumably dissolved post-settlement. Genetic data from Korean samples suggest the range extends from the Aleutians into the Northwest Pacific as well (see remarks). The presence of this species in California is unconfirmed; if present, it is likely only present in deep water (>100 m; see remarks).

Remarks. This species was described from samples collected near Vancouver Island (Pacific Canada), but the morphological similarity to other Suberites has led to uncertainty regarding its species status (Austin et al. 2014) . A recent revision acknowledged the difficulty of morphological identification in this group, but supported the concept of S. latus as a North Pacific relative of the North Atlantic S. ficus complex, and went on to propose that "this species complex would be a good candidate for DNA barcoding" (Austin et al. 2014) . Here we present DNA data from samples collected in British Columbia, near the type location. These samples are genetically differentiated from morphologically similar Atlantic species such as S. ficus Johnson, 1842 andS. pagurorum Solé-Cava & Thorpe, 1986, supporting species status for S. latus. It is also notable that the microstrongyles of this species are more common in the choanosome, which contrasts with other Suberites, where they are more common in the ectosome.

We were unable to confirm the presence of S. latus in California. All freshly collected and previously vouchered material from California's shallow waters was found to be either S. kumeyaay sp. nov. or S. lambei. Consistent with the lack of vouchered samples from shallow water in California, amateur observations of Suberites with hermit crabs on the site iNaturalist extend from Kodiak, Alaska to the San Juan Islands in northern Washington, but are absent farther South. Suberites associated with hermit crabs are found in deep water in Southern California, by environmental monitoring projects, nearly always below 100 m (Megan Lilly, City of San Diego, pers. comm.). Well-characterized samples are morphologically differentiated from the Northern S. latus, and we allocate them to Suberites californiana sp. nov., described below. However, it remains possible that S. latus also ranges into California in deep water, as few deepwater California samples are available for review.

At the cox1 locus, Genbank records from Korea, previously identified as S. ficus, are genetically identical to S. latus, raising the possibility that the species range extends from the Aleutians into the Northwest Pacific (figure 2). Confirmation would require DNA sequence from additional loci and morphological data, as the slow rate of evolution at cox1 in sponges sometimes leads to identical sequences between populations that are good species under most species concepts (Huang et al. 2008; López-Legentil et al. 2010; Pöppe et al. 2010; Turner & Pankey 2023) . Shallow samples from Peru were recently described as Suberites cf. latus (Cóndor-Luján et al. 2023 ). Due to the disjunct distribution, and lack of association with hermit crabs, we think it likely that DNA data would reveal these to be a different species.

Suberites californiana sp. nov. Turner 2023

Material examined. Holotype: P362, Near Anacapa Island, (34.00, -119.39), 223-242 m, 4/3/52.

Material examined. Holotype: TLT1439, San Diego Bay, 14.25 m, 8/31/23; Paratypes: TLT1340, Offshore San Diego, (32.53197, -117.18800), 31 m, 7/25/17; TLT1339, San Diego Bay, 10 m, 8/17/98; TLT1341, San Diego Bay, 10 m, 8/13/98; TLT1441, San Diego Bay, 10 m, 8/14/98; TLT1442, San Diego Bay, 10 m, 8/24/98.

Etymology. Named for the Kumeyaay people, who have inhabited the San Diego area since preColumbian times.

Morphology. Massive, amorphous, lobate sponges. Bright orange to brownish orange alive, light brown preserved. The holotype is a roughly triangular lobe cut from a larger sponge, 9 x 7 x 2 cm across. It has scattered oscula, 1-5 mm in diameter. Dried samples available for examination measure up to 22 x 10 cm, but field photos indicate larger sponges are sometimes found. The holotype was preserved in 95% ethanol, but other samples were preserved in formalin and/or dried.

Skeleton. Ectosomal skeleton of tylostyles in upright bouquets (figure 5D), with microstrongyles among the bouquets. Choanosomal skeleton of tylostyles in confusion with no apparent pattern. A few microstrongyles are found scattered in the choanosome, but are much less abundant than at sponge surface.

Spicules. Long and short tylostyles, with spined centrotylote strongyles as microscleres. Tylostyles: usually slightly curved and weakly fusiform (thickest in center), but longest tylostyles are often straight. Heads are typically round and terminal, but frequently off-center and sometimes sub-terminal (figure 5E). The length distribution is bimodal, but with many styles of intermediate size that make it difficult to categorize spicules as short vs. long. Choanosomal styles average 60% longer than ectosomal styles.

Holotype. Ectosome, 12131933272 x 537310 ¿m (n=40); choanosome, 15132603305 x 13739 ¿m (n=40); both domains combined, 12132263305 x 137310 ¿m (n=80).

All samples pooled, 8132233324 x 137313 ¿m (n=338), with modes at 135 and 285 ¿m. Centrotylote microstrongyles: microspined, but appearing smooth under light microscopy. Swollen center variable from well-developed to barely present (figure 5C); it is generally in the center but sometimes off-center or at one end. Spicules usually slightly curved but sometimes straight. Holotype: 13329345 x 13233 ¿m (n=97 length, n=35 width). All samples pooled: 133 28345 x 13234 ¿m (n=175 length, n=112 width).

Distribution and habitat. Common in the shallow waters of San Diego Bay in Southern California. One sample was also found 6 km offshore of San Diego in 31 m of water. Remarks. Environmental monitoring projects in San Diego Bay, in extreme southern California, frequently yield large Suberites that are not associated with gastropod shells or hermit crabs; these were previously referred to as "Porifera sp. SD 4" by the Southern California Association of Marine Invertebrate Taxonomists (Megan Lilly, City of San Diego, pers. comm.). These differ from Suberites californiana sp. nov. in that they are 1) not associated with gastropod shells and hermit crabs, 2) found in shallow water (4331 m, vs. 1803242 m foSr. californiana), 3) have much shorter tylostyles (modes at 135 and 285 ¿m vs. 250 and 600 ¿m for S. californiana; maximum lengths 3053324 ¿m vs. 6003693 ¿m forS. californiana), and 4) possess microstrongyles that are slightly, but significantly (p<0.001) longer than S. californiana (mean = 28 ¿m vs. 23 ¿m for S. californiana). Suberites kumeyaay sp. nov. is also morphologically distinguished from S. latus in that they are 1) not associated with gastropod shells and hermit crabs, 2) have abundant ectosomal microscleres (vs. rare and/or choanosomal in S. latus), 3) have slightly shorter tylostyles (modes at 135 and 285 ¿m vs. 170 and 310 ¿m for S. latus; maximum lengths 3053324 ¿m vs. 3543409 ¿m fSo.rlatus), and 4) possess microstrongyles that are slightly, but significantly (p<0.001) longer than S. latus (mean = 28 ¿m vs. 24 ¿m forS. latus). It is easily differentiated from all other Suberites in the region, which lack microscleres.

Previously collected samples of S. kumeyaay sp. nov. were either preserved in formalin or dried, but we were able to generate 28S sequence from a freshly collected sample, preserved in ethanol, graciously collected for this purpose by Brandon L. Stidum. This 28S sequence is most similar to S. ficus, a North Atlantic species that is very morphologically similar to S. kumeyaay sp. nov. These species are quite genetically divergent, however, with 3.60% pairwise sequence differences, which is much higher than the sympatric, reproductively isolated S. ficus and S. pagorum (0.84% divergence). Combined with their highly disjunct distribution, this leads to confidence in assigning Southern California samples a new species.

Suberites sp. (Bakus & Green 1987; Hartman 1975; Lee et al. 2007) Material examined. Holotype: RBC982-00066-001, Barkley Sound, George Frasier Island, British Columbia, (48.90833,-125.51333), intertidal, 7/29/76; Paratype: RBC982-00069-001, Gowlland Pt., S. Pender Island, British Columbia, (48.73667,-123.18167), 3 m, 3/18/77; CASIZ302850, South Shore, San Nicolas Island, San Nicolas Is., (33.21717,-119.47450), 10 m, 9/28/11; CASIZ302796, The Arch, San Clemente Island, San Clemente Is., (32.96000,118.51533), 7 m, 9/19/11; TLT192, Arroyo Hondo, Santa Barbara, (34.47182,-120.14262), 4-8 m, 7/29/19; TLT441, Elwood Reef, Santa Barbara, (34.41775,-119.90150), 9-15 m, 10/23/19; TLT374, Wreck of the Cuba, San Miguel Is., (34.03009,-120.45575), 6-12 m, 10/6/19; BULA0547/NHMLA16557, Halfway Reef, Los Angeles, (33.76265,-118.42560), 15-23 m, 8/23/19; BULA0549/NHMLA16559, Halfway Reef, Los Angeles, (33.76265,-118.42560), 15-23 m, 8/23/19; TLT432, Elwood Reef, Santa Barbara, (34.41775,-119.90150), 9-15 m, 10/23/19; TLT672, La Jolla Cove, San Diego, (32.85227,-117.27239), 15 m, 8/14/20; TLT684, Hazard Canyon, San Luis Obispo, (35.28959,-120.88415), intertidal, 12/12/20; TLT660, Hazard Canyon, San Luis Obispo, (35.28959,-120.88415), intertidal, 12/12/20; TLT825, Cave Landings, San Luis Obispo, (35.17535,-120.72240), intertidal, 2/6/21; TLT1270, Zebra cove, Anacapa Is., (34.01000,-119.44000), 6-13 m, 6/17/22; TLT806, Three sisters, Santa Cruz Is., (33.98675,119.58884), 6-13 m, 4/16/21; TLT991, Farnsworth Bank, Catalina Is., (33.34380,-118.51650), 21-27 m, 7/10/21; TLT1102, CRABS, Monterey, (36.55377,-121.93840), 10-17 m, 9/21/21; BHAK-03707, Simon`s Landing Drop-down, Tatoosh Island, Washington, (48.391917, 124.733236), intertidal, 6/15/18; BHAK-03699, Strawberry Draw Overhang, Tatoosh Island, Washington, (48.392003, -124.737667), intertidal, 6/15/18; UF4317, Cape Arago, south Middle Cove, Oregon, (43.30477,-124.40061), intertidal, 6/5/19; UF4300, Cape Arago, Middle Cove, Oregon, (43.30477,-124.40061), undetermined, 6/4/19; BHAK-02365, West Beach, Calvert Island, British Columbia, (51.6495,-128.155), 5 m, 8/2/17; TLT1395, Cape Alava, Washington, (48.17134, -124.75227), 4-12 m, 7/29/23; TLT1386, Cape Alava, Washington, (48.17134, 124.75227), 4-12 m, 7/29/23; TLT1313, Cape Alava, Washington, (48.17134, -124.75227), 4-12 m, 7/29/23; TLT1426, Tatoosh Island, Washington, (48.39370, -124.73305), 4-11 m, 7/31/23. Morphology. Cushion shaped to massive and asymmetrical. Firm and rubbery with velvety surface. Highly contractile; becomes a firm mass without apparent oscula when exposed by the tide. When alive and relaxed, subtidal samples display a partially transparent network of pores on the surface and large, obvious oscula on raised chimneys. Samples vary from yellow to brownish-yellow alive; beige after preservation.

Skeleton. Ectosomal skeleton of tylostyles in upright bouquets. Tylostyles in choanosome in confusion with no apparent pattern.

Spicules. Tylostyles only. Often straight, but some samples have a high frequency of spicules with slightly curved, bent or sinuous shapes. Heads of California samples are frequently lumpy and uneven (figure 4E), but this was rare in samples from farther North (e.g. British Columbia, Washington). Most samples display a uniform thickness until reaching a sharply tapering point, but some samples slightly fusiform (thickest in center) and/or more gradually tapering. Slightly shorter in the ectosome.

For one sample (TLT296), spicules were measured separately for the ectosome (19934293559 x 437311 ¿m (n=124)) and choanosome (26134793599 x 337312 ¿m (n=91)) and found to be significantly smaller in the ectosome (Wilcox p-value = 2.6e-07). The combined distribution was not distinctly bimodal due to the similarity of the distributions.

Spicules were measured from 15 California samples, collected from Monterey Bay in Central California to La Jolla cove in Southern California, including 6 different Channel Islands. All samples together: 14834183615 x 237318 ¿m (n=570). Mean values varied greatly across samples, with mean lengths from 282 to 511 ¿m and mean widths from 5 to 12 ¿m. Spicules were measured from the holotype and a paratype from British Columbia, and a sample from Tatoosh Island, Washington. Mean lengths of these samples were similar to California samples (387 to 483 ¿m) but mean widths were slightly larger than the largest California sampels (13 to 16 ¿m).

All measured samples combined: 14834263615 x 239326 ¿m (n=748).

Distribution and habitat. Known from Queen Charlotte Island in British Columbia to San Diego, CA, with a depth range from the low intertidal to 30 m. Common in the intertidal in northern and central California and probably farther North; southern-most intertidal samples known from Malibu, California. Presence-absence surveys in subtidal kelp-forest habitat in California suggest it is widespread but uncommon; it was seen at 10% of investigated reefs, evenly spread from Monterey to San Diego counties (the entire investigated range). The nudibranch Doris montereyensis is frequently observed feeding on this species.

Remarks. A species of Suberites was first noted in the Central California intertidal nearly 50 years ago, but despite being locally abundant, it has remained unnamed (Hartman 1975; Lee et al. 2007) . It is fairly common in the intertidal, especially in northern and central California; as such, it is surprising that it was not noted during the extensive intertidal work of de Laubenfels in the 1920s (de Laubenfels 1932). It is possible that de Laubenfels' subtidal Ficulina suberea lata were in fact a mix of S. latus and S. lambei, but all of those samples were subtidal or in beach wrack rather than intertidal. Consistent with this possibility, de Laubenfels indicated lumpy tylostyle heads in these samples, which were common in our California samples of S. lambei (figure 25, de Laubenfels 1932) but not seen in other Suberites species we examined.

Suberites lambei was described from British Columbia, with the thickness of the spicules as a key trait distinguishing it from S. latus (Austin et al. 2014) . Austinet al. compared their new species to a few samples from Central California, but were skeptical that they were the same species because 1) most California samples had much thinner spicules and 2) the populations were disjunct, as they were unaware of samples between British Columbia and Central California. Here, we have collected additional samples from British Columbia, Washington, Oregon, Central California, and Southern California, from both the intertidal and the subtidal. The spicular dimensions of these samples vary greatly, especially in width, but the continuous distribution of values suggests this is intraspecific variation. DNA data at both cox1 and 28S support this assertion (figures 1-2). We therefore have no evidence supporting more than one species among these samples, and assign them all to S. lambei.

This species was previously distinguished from S. latus by spicule width, but we show that this character is variable. Suberites lambei never have microscleres, which differentiated them from all S. kumeyaay, all S. californianai, and nearly all S. latus. As some samples of S. latus may lack microscleres, these samples can be morphologically distinuished in the following ways. Suberites latus is often orange (S. lambei is yellow or light brown); S. latus is usually associated with hermit crabs (not seen in S. lambei); S. latus has a substantial difference in tylostyle length between ectosome and choanosome4 nearly twice as long in the choanosome 4 leading to a strongly bimodal distribution when they are combined (not seen in S. lambei). The tylostyles are also usually shorter in S. latus than S. lambei, but the shortest S. lambei sample we investigated was similar to S. latus (genotyping confirmed this sample was S. lambei), so this character is insufficient in isolation. Finally, we note that these species are differentiated by depth range in California, but not in more Northern parts of their shared range.

Etymology. Named for its resemblance to cultivated "button" mushrooms in the genus Agaricus. Morphology. Roughly spherical, 14 mm in diameter. Numerous small oscula (~1 mm in diameter) flush with surface in living specimen; constricted in preserved specimen. The ventral side appears to have been attached to the substrate over a smaller surface than the entire diameter of the sponge, leading to an incipient stalked appearance, and making the overall morphology similar to the cap of the cultivated "button" mushroom. Pale yellow alive, white when preserved. Cross-section of the preserved sample reveals a visually distinct ectosomal layer 0.2-0.5 mm in thickness (figure 5D). Microscopically hispid.

Skeleton. Upright bouquets of tylostyles in the ectosome, with tips protruding 100-150 ¿m beyond sponge surface. Bouquets are supported by multispicular columns of tylostyles, creating a somewhat radial architecture, becoming increasingly confused towards the interior of the sponge.

Spicules. Large and small tylostyles, with overlapping size distributions. The distribution of tylostyle length is bimodal when all tylostyles are considered together, with modes at approximately 350 ¿m and 1050 ¿m. If 700 ¿m is used as a dividing point between classes, then the choanosome contains only long tylostyles, while the ectosome contains 20% long tylostyles and 80% short tylostyles. All tylostyles are straight or slightly curved, usually with symmetrical, round heads, but some heads are sub-terminal. Most spicules are thickest near head and gently tapering throughout their length, but some are slightly fusiform. The gradual tapering differentiates this species from the other Suberites found in California, but R. gadus is similar. All spicules together: 204359831335 x 6311324 ¿m (n=161), modes at approximately 350 ¿m and 1050 ¿m.

Choanosome only: 8603107431335 x 11317324 ¿m (n=23) Ectosome only: 223354831073 x 639316 ¿m (n=87) Only spicules < 700 ¿m in length: 20434093692 x 639316 ¿m (n=107) Distribution and habitat. Only seen at a single location to date: the kelp forest off of Point Loma, near the southern border of California. Numerous small, roughly spherical sponges matching the morphology of the sampled individual were seen at this location, which were all likely this species.

Remarks. This species could arguably be placed in either Suberites or Rhizaxinella. The body shape and somewhat radial skeleton are more consistent with Rhizaxinella, but sponges in this genus are carried on a long, often branching stalks, which are missing in this species. With the added evidence of a close genetic relationship to other Suberites, we feel this placement is a better fit. It is likely that the spherical body of the sponge has led to some convergent evolution of the skeleton from confused into a more radial architecture.

This species is quite closely related to S. lambei, with only 0.3% sequence divergence at the 28S locus. Divergence at the cox1 locus is higher (1.7%), and with consistent placement and both loci and very distinct spicule morphology, it is confidently assigned to a separate species. As discussed in the S. lambei section, tylostyles in that species have a unimodal length distribution, with average values between 282 ¿m and 499 ¿m per sample. We measured 750 spicules across 17 individuals throughout the range of S. lambei (including a sample from San Diego, near where S. agaricus was collected) and found the maximum spicule length to be 615 ¿m. This is in stark contrast to S. agaricus, where minimum sizes of the long tylostyles are longer than this maximum value for S. lambei. These long tylostyles also set the new species apart from all other Suberites in the region: S. latus, S. kumeyaay sp. nov., S. californiana sp. nov., S. concinnus Lambe, 1895 (which lacks tyles), and S. mineri de Laubenfels, 1935 (which is cup-shaped). Suberites baffini ( Brøndsted 1933 ) has spicules of similar length to S. agaricus, and also has a roughly radial architecture. These are unlikely to be conspecific, however, as S. baffini is described as having prominent spicular fibers, whereas no spongin or other fiber is present in S. agaricus; moreover, S. baffini is known from deep water (1200 m) the far North Atlantic (Baffin Bay, Canada).

The spicules of the new species are similar to Rhizaxinella gadus. The spicules of R. gadus are significantly longer (Wilcox rank-sum test p < 0.001), but examining more individuals may reveal overlap in size distributions. As described in the R. gadus section below, these species can be differentiated based on skeletal organization, the lack of a branching stalk in S. agaricus, and DNA sequence; they also likely have depth range differences, but further work could reveal overlap in depth range.

Figures 1-2, 8

Suberites gadus (de Laubenfels 1926, 1932) Material examined. SIO-BIC P1999, San Pedro Basin, whale fall, (33.5736, -118.43676), 885 m, 8/3/21; P2000, San Pedro Basin,, (33.57650, -118.43484), 882 m, 8/3/21; SIO-BIC P2015, barrel 1, San Pedro Basin, DDT barrel dump site, (33.56602, -118.42426), 885 m, 8/4/21; SIOBIC P2016, barrel 2, San Pedro Basin, DDT barrel dump site, (33.56607, 118.42566), 885 m, 8/4/21. All collected by the Schmidt Ocean Institute using the ROVSuBastian.

Morphology. Holotype previously described as cylindrical and peanut-shaped sponges atop branching stalks 1 cm thick. Only small fragments of newly analyzed samples were available for morphological analysis, but ROV video from their collection shows globose sponges 2-6 cm in diameter; many were atop thick and sometimes branching stalks, though stalks were not apparent on all sponges. Oscula sometimes visible either as a few scattered holes near the apex, flush with sponge surface, or as a larger single opening, also at apex (oscula not visible in measured samples). White to pale yellow alive; white preserved. Microscopically hispid.

Skeleton. Short tylostyles in upright bouquets and bundles in the ectosome, with tips protruding ~300 ¿m beyond the sponge surface. Long tylostyles occurring both singly and in bundles of up to 20 spicules in the choanosome. Single spicules and spicule bundles are both found in confusion with no apparent spicule tracts. Samples available for examination did not include stalks; stalks previously described as having a "central axis of densely packed longitudinally placed spicules" (de Laubenfels, 1932) and an ectosomal skeleton the same as the lobes.

Spicules. Long and short tylostyles. Most with well-formed, round, terminal heads, but occasionally with subterminal heads or both. Straight or slightly curved; most are gradually tapering throughout their length, but some taper for only half their length or less. Long tylostyles: nearly all the spicules in the choanosome are in the longer size class, while a minority of spicules in ectosomal spicule preps are also in the larger size class. Distributions are overlapping; if 850 ¿m is used as the dividing point, then long tylostyles in SIO-BIC P1999 are 8523107631512 x 9312317 ¿m (n=74); long tylostyles of SIO-BIC P2016 are 9073133831734 x 8312317 ¿m (n=34).

Short tylostyles: Only slightly thinner than long tylostyles, and otherwise with the same shape. Using 850 ¿m as the dividing point between long and short tylostyles, short tylostyles in SIOBIC P1999 are 22935213817 x 538312 ¿m (n=77); short tylostyles of SIO-BIC P2016 are 2603 4623776 x 437310 ¿m (n=57).

All tylostyles of SIO-BIC P1999, considered together: 229379331512 x 5310317 ¿m (n=151). Distribution is bimodal, with modes at 350 and 1025 ¿m. Tylostyles of SIO-BIC P2016: 2603 78931734 x 439317 ¿m (n=91), with modes at 350 and 1300.

Distribution and habitat. The holotype was found near Pacific Grove in Central California; it was tangled in fishing gear and sampled at uncertain depth (depth first reported with skepticism as 180 m and later as 30 m (de Laubenfels 1926, 1932)). Newly collected samples were found on whale falls and on and near metal barrels in the San Pedro Basin of Los Angeles, 880-890 m depth. Some protruded from sediment, with the branching portions of the stalk within the sediment. This species may be common in deep water in the Northeast Pacific, but further samples are needed for confirmation (see remarks).

Remarks. This species was described from a single sample found near Pacific Grove, California. Newly collected samples analyzed here match the described sample very well in terms of spicule sizes, spicule morphology, and gross morphology.

There are several hundred records of this species in the Global Biodiversity Information Facility (gbif.org), but nearly all are from the NOAA Deep Sea Corals Research and Technology Program. These records are based on the gross morphology of samples seen on video transects, and were not collected and examined, so they should be considered tentative. We examined the only other sample in the database that was collected in California (CASIZ165849), and found that it did not match the description of R. gadus (the sample contained multiple size classes of tylostyles, but they were smaller than R. gadus, and the sample also had abundant strongyloxea up to 2500 ¿m long; only a small fragment was available, so we were unable to determine the identity of this sample, but it is likely from an undescribed deep-water species that may also be in the Suberitidae). It therefore remains possible that there are multiple species of stalked sponges matching the morphology of R. gadus in the region, but this remains to be confirmed. The remaining GBIF records are from Mexican waters and were not examined here.

Pseudosuberites latke sp. nov. Turner 2023

Figures 1-2, 9 Material examined. Holotype: TLT1131, Fire Rock, Pescadero Point, (36.55898, -121.95110), 10-22 m, 8/10/21. Paratypes: TLT538, Goalpost, Point Loma, (32.69438, -117.26860), 12-15 m, 2/8/20; TLT585, Train wheels, Point Loma, (32.65205, -117.26243), 9/19/20; 15-29 m, 9/19/20; TLT1143A, Tanker Reef, Monterey Bay, (36.60607, -121.88160), 10-16 m, 8/9/21; TLT1128, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 9/8/21; TLT19, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 4/17/19; TLT433, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 10/23/19.

Morphology. Holotype is a bushy mass, 30 mm thick 8 mm across, composed of a reticulate network of individual branches 2-5 mm in diameter. Other samples similar, but some encrusting instead of bushy, 5 mm thick, with the reticulate mass prone on substrate rather than upright.

Bright yellow alive, white preserved. Oscula sometimes visible in photos of living sponges.

Skeleton. A dense confused mass of tylostyles is formed into a reticulate network. Plumose bundles and individual spicules protrude from this confused mass and pierce the surface of the sponge, making ectosome difficult to detach. Ectosomal sections contain tangential tylostyles in confusion.

Spicules. Tylostyles only. Highly variable in length and width, but without bimodal distributions. Nearly all with well-formed, round, terminal heads, but occasionally with subterminal heads. Straight or slightly curved, with uniform thickness along shaft until reaching a sharply tapering point.

Holotype: 24734003622 x 6312323 ¿m (n=131). When ectosome and choanosome are prepared separately, spicules in ectosome are significantly shorter (p=0.016), but distributions are broadly overlapping (ectosome only: 24733613473 x 6310315 ¿m (n=31); choanosome only: 26334103 622 x 6313323 ¿m (n=42)).

All samples combined: 17633883730 x 2311323 ¿m (n=265). Means of each sample vary from 341 to 420 ¿m in length, 6 to 12 ¿m in width.

Distribution and habitat. All known samples were recently collected by one of the authors. Presence-absence surveys in subtidal kelp-forest habitat in California found this species to be uncommon from Monterey to San Diego (the entire investigated range); the shallowest samples were found at 10 m and the deepest at 30 m (the maximum depth investigated). It was most abundant in southern San Diego County, where it was seen at 5/12 sites investigated. Many of these were deeper (20-30 m) sites, where it may be more abundant than in very shallow water. Not seen at any island sites, intertidal sites, or on human structures. The nudibranch Doris montereyensis has been observed feeding on this species.

Remarks. Suberites and Pseudosuberites are currently distinguished by their ectosomal skeleton (a tangential and usually detachable crust carried by subectosomal spicule brushes in Pseudosuberites, a palisade of small tylostyles in Suberites) and their tylostyle size distributions (a smaller ectosomal size class in Suberites and no localized size classes in Pseudosuberites). This species did have some tangential ectosomal spicules, though the surface was not easily detachable as it is in some Pseudosuberites. Two size classes vs. one size class of tylostyles is likely not a discrete character, as species like P. latke sp. nov. and S. lambei have no size classes but significantly smaller spicules in the ectosome. Combining DNA and morphological data leads us to assign this species to Pseudosuberites. It is very distantly related to all genotyped Suberites, but forms a clade with Pseudosuberites from South Africa, Antarctica, and Korea at the cox1 locus.

The dense reticulation of confused tylostyles makes this species very distinct from any other species known from the region. No other Pseudosuberites are described from the Northeast Pacific. The type species P. hyalinus was once said to have been collected in Pacific Mexico (Dickinson 1945) ; little description is provided, but the spicules are listed as three timesas long as those of P. latke sp. nov.. A more recent publication on the sponges of the tropical Mexican Pacific includes a "Pseudosubertes sp." but no morphological description is yet available (Carballo et al. 2019) .

Protosuberites sisyrnus (de Laubenfels 1930) Material examined. Fragment of holotype, found in the Natural History Museum Los Angeles collection, labeled as "21413 (Part of type)", with collection information that fits holotype. The holotype is also stored in the Smithsonian as USNM21413. Collected South of the breakwater, San Pedro, California, 45 m, 4/5/1924.

Morphology. Sponge is encrusting on a tangled mass of polychaete tubes and other debris. Visibly hispid. Difficult to further characterize due to the age and complexity of the sample.

Light brown post-preservation.

Skeleton. Not investigated; previously described as erect tylostyles in the ectosome and few spicules, in confusion, in the choanosome.

Spicules. Tylostyles; both straight and slightly curved spicules are common, most with round terminal heads, but occasionally subterminal. Consistent thickness through shaft until reaching a sharply tapering point. Considerable variation in length and width but distribution unimodal. 24633863574 x 7312319 ¿m (n=50) Distribution and habitat. Described from two samples dredged from 45-54 m near Los Angeles and Catalina Island in Southern California. Museum records indicate occurrences in British Columbia, but this remains to be confirmed, as samples we examined were misidentified (see remarks).

Remarks. We were unable to locate any samples of this species in the field (previously collected samples were from a depth range we were not able to investigate). Vouchers previously identified as Protosuberites were loaned to us from British Columbia, but these proved to be misidentified (they are likely in the Microcionidae, as spicules consist of palmate chelae, toxas, and two sizes of acanthostyles). We attempted to sequence DNA from several loci from the holotype fragment in the Natural History Museum of Los Angeles but were unsuccessful. The quantitative spicule data presented here should aid in future attempts to locate and characterize this species.

Protosuberites sp.

Material examined. TLT443, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 10/23/19.

Morphology. Thinly encrusting on a small (~1 cm) blade of red algae. Bright yellow alive. Skeleton. Not investigated; the entire sample was consumed for DNA extraction and spicule preparation.

Spicules. Tylostyles, straight or weakly curved. Often with well-formed terminal heads, but a high frequency of heads that are sub-terminal or slightly lumpy. Longer spicules are gently tapering for much of their length, but shorter spicules have a consistent width until tapering sharply at tip. Highly variable in length, but distribution is unimodal. 10332463474 x 23437 ¿m (n=43).

Distribution and habitat. Found at a single location, Elwood Reef, a shallow, rocky reef with kelp forest cover near Santa Barbara, California.

Remarks. Thinly encrusting Suberitidae are currently placed in either Terpios or Protosuberites (Morrow & Cárdenas 2015; van Soest 2002) . Though we did not examine the skeleton of this small sponge, we think it is more likely to be in Protosuberites due to the shape of the tylostyles and the lack of gelatinous consistency. DNA supports this placement, as the species falls into a clade containing a worldwide sample of other Protosuberites.

With only a single sample that was consumed without investigating the skeleton, and noting how variable we found spicule lengths to be in S. lambei, we conservatively leave this sample as an unknown for the time being.

Homaxinella amphispicula (de Laubenfels 1961) Figures 1-2, 11

Syringella amphispicula (de Laubenfels 1961) Material examined. TLT1137, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 9/8/21; TLT1177, Elwood Reef, Santa Barbara, (34.41775, -119.90150), 9-15 m, 9/1/21; RBC018-00132-001, Sechelt Peninsula; Sakinaw Lake, British Columbia, (49.56760, 123.80415), 0-15 m, 7/23/15.

Morphology. Upright, branches 2-10 mm in diameter split and anastomose to form a bush mass. Occasional oscula visible at apex of branches in living specimen; contracted and invisible after collection. Dull brownish-yellow alive, beige preserved.

Skeleton. Spicule bundles run parallel to branches in the core of each branch. additional bundles, and single spicules in confusion, radiate out from core to periphery. Upright spicules pierce the surface at the periphery in disorganized bouquets and individually.

Spicules. Long thin styles, slightly curved or rarely strongly curved or sinuous; heads vary from simple styles to subtylostyles, with intermediates. Highly variable in length, but not falling into easily delineated size classes. Significantly smaller in ectosome, with longest styles (>515 ¿m) not found in ectosome. Both California samples also had oxeas, but they were rare (<1% of spicules).

Styles of California samples: 19734223737 x 236312 ¿m (n=171); ectosome only of sample TLT1177: 22433383513 x 23437 ¿m (n=44).

British Columbia sample with slightly longer and much wider styles: 28835253733 x 6312317 ¿m (n=38).

Oxeas of California samples: 35735363687 x 33537 ¿m (n=12).

Distribution and habitat. Previously known from the San Juan Islands in Northern Washington to British Columbia, 10-70 m in depth (Austin et al. 2014) . New samples from Santa Barbara California are a Southern range extension of over 1,500 km.

Remarks. As this species was previously known only from British Columbia and the San Juan Islands, we were surprised to find it in Santa Barbara. The first sample was collected by Christoph Pierre and Christian Orsini (UCSB), and a second sample was collected soon after, at the same location, by one of us (T. Turner). Both California samples have rare oxeas, which were not found in the large number of British Columbia samples previously analyzed (Austin et al. 2014) , nor were we able to find them in the British Columbia sample we analyzed. California samples also have much thinner spicules than British Columbia samples (but we note that thickness is variable in other species, and is sometimes associated with water temperature or chemistry). We were able to generate DNA data at the 28S locus for one British Columbia sample and one Santa Barbara sample, which differ by 0.3%; these samples differ from the North Atlantic H. subdola (Bowerbank 1866) by 0.8%. These limited DNA data are therefore consistent with both conspecificity or closely related taxa. As the morphological differences are slight, we conservatively assign the California samples to H. amphispicula, but note that further work may reveal them to be a different species. 4A. Globular sponges atop branching stalks; choanosomal skeleton consisting of spicules and spicule bundles in confusion; collected in deep water: Rhizaxinella gadus (de Laubenfels 1926) 4B. Sponge unstalked; choanosomal skeleton containing multispicular columns; collected in shallow water: Suberites agari cus sp. nov. Turner 2023 5A. Choanosomal skeleton a dense, confused mass of tylostyles formed into a reticulate network; ectosome with only sparse brushes of upright tylostyles and some tangential tylostyles: Pseudosuberites latke sp. nov. Turner 2023 5B. Choanosomal skeleton confused and/or containing radial tracts; ectosome packed with numerous upright tylostyles in a palisade or in bouquets ...6 6A.

1A.