Argonauta nouryi Lorois, 1852

Finn, Julian K., 2018, Recognising variability in the shells of argonauts (Cephalopoda: Argonautidae): the key to resolving the taxonomy of the family, Memoirs of Museum Victoria 77, pp. 63-104 : 65-77

Argonauta nouryi Lorois, 1852 ; the A. nouryi / cornutus complex

In spring each year, small argonauts wash up in large numbers on beaches in the southern Gulf of California (Gonzales-Peralta in Saul and Stadum 2005). These small argonauts are regularly attributed to two species: A. nouryi Lorois, 1852 and A. cornutus Conrad , 18541 ( Garcia-Dominguez and Castro-Aguirre 1991; Gonzales-Peralta 2006).

A. nouryi was described by Lorois in 1852. The identification of this species resides solely in features of the shell, which is described as elliptical with numerous fine lateral ribs and weak keel tubercles. Fig. 3 incorporates a

1 A third large form also washes up on southern Gulf of California beaches in spring and is regularly attributed to the species A. pacificus , a synonym of A. argo ; see Finn (2013) for details.

reproduction of the illustration presented by Lorois, 1852 (plate 1, figure 5), and illustrations of a shell from the Gulf of California that is consistent with the original description (shell #109, SBMNH 345768). According to Keen (1971) “the ‘shell’ is more elliptical than that of A. cornutus , with only the early part of the coil moderately well tinged with brown along the wide and weak tuberculate keel. The surface is delicately ribbed and has a finely granular texture” (p. 895). Voss (1971) believed that “ Argonauta nouryi is a distinctive species […]. The shells are longer than in any other species of Argonauta , the ribs are more numerous, there are no distinct tubercles marking the edges of the carinal area; the carina is wide, very convex, and covered by numerous, small, blunt tubercles formed by the crisscrossing of the ribs” (p. 32).

Argonauta cornutus was described by Conrad in 1854. The identification of this species also resides solely in features of the shell, which is described as having a broad keel, large keel tubercles and large ears. Fig. 4 incorporates a reproduction of the illustration presented by Conrad, 1854 (plate 34, figure 2), photographs of the type specimen (ANSP 63496) and illustrations of a shell from the Gulf of California that is consistent with the original description (shell #74, SBMNH 345766). According to Keen (1971), “the surface of the yellowish-white ‘shell’ is finely granular, the spines and part of the spire dark brown, the keel relatively broad, and the two long axial expansions suffused with purplish brown” (p. 894). Voss (1971) summarised that “ Argonauta cornutus seems best characterised by the few radial ribs, the presence of fine sharp tubercles or papillae over the sides of the shell, the few rather sharp, large carinal tubercles on each side, the convex carinal surface, and the few, large, blunt tubercles on the carinal surface between the two rows of carinal boundary tubercles” (p. 32).

The distributions of these two species are reported to overlap, with A. cornutus known from the Gulf of California to Panama and A. nouryi being widespread in the equatorial Pacific, ranging from the west coast of Southern California to Peru ( Keen 1971).

A mixed lot

As described in the Materials and Methods section above, the 157 shells in the collection at SBMNH were collected on the same beach in Baja California on the same day. These shells had previously been identified as representing both A. cornutus and A. nouryi and were registered accordingly: SBMNH 345766, Argonauta cornutus 93 shells; SBMNH 345768, Argonauta nouryi , 64 shells.

Initial examination of the lots indicated that the shells had been attributed to either A. cornutus or A. nouryi based on the presence or absence of ears – a character historically attributed to only A. cornutus . Further examination of the lot revealed that separation of the shells into two distinct groups (i.e. either A. cornutus or A. nouryi ) was not as straightforward as first thought. While some shells within the lot displayed all the characters associated with either A. cornutus or A. nouryi , the lot also appeared to contain shells with combinations of the attributes of the two shell types. To illustrate this variation, three shells of similar size but varied appearance were selected. Fig. 5 presents photographs of these three shells from multiple perspectives:

•

•

• Shell #74 (SBMNH 345766), cornutus-type voucher (fig. 5a, i–iv and fig. 4c). Shell morphometrics: ShL 65.0; ShW 4.0; ShB 40.7; RC 45; EW 58.1; ApL 45.9; ApW 28.4; KW 15.6; KTC 27.

Shell #42 (SBMNH 345766), intermediate voucher (fig. 5b, i–iv). Shell morphometrics: ShL 61.2; ShW 3.1; ShB 36.4; RC 47; EW 36.1; ApL 43.1; ApW 30.9; KW 14.0; KTC 32.

Shell #109 (SBMNH 345768), nouryi-type voucher (fig. 5c, i–iv and fig. 3b). Shell morphometrics: ShL 66.5; ShW 2.4; ShB 39.9; RC 54; EW (28.3); ApL 48.8; ApW 32.5; KW 15.8; KTC 54.

While it would be straightforward to attribute shell #74 (fig. 5a) to A. cornutus Conrad, 1854 , and shell #109 (fig. 5c) to A. nouryi Lorois, 1852 , placement of shell #42 (fig. 5b) presents problems. While shell #42 possesses the aperture shape of A. cornutus , it lacks its protruding ears. While shell #42 possesses the keel tuberculation and reduced ventral keel tubercles of A. nouryi , its dorsal keel tubercles are large and pronounced.

To determine whether there was a significant difference between eared and earless shells within the lot, a quantitative approach was undertaken. All intact shells within the lot were individually measured and weighed. All shells were designated as being either eared or earless based on the relative EW and ApW measurements. Because EW is an external measurement (i.e. measured across the extremities of the opposing ears) and ApW is an internal measurement (i.e. measured between the lateral walls of the shell), 1.0 mm was added to the ApW to accommodate for the thickness of the lateral walls of the shell. Shells were classified as follows:

•

• eared = EW> ApW + 1.0 mm (103 shells)

earless = EW ≤ Apw + 1.0 mm (35 shells).

Scatter plots were generated to compare eared and earless shells for all measured characters. Characters of primary interest were those previously reported to distinguish A. cornutus and A. nouryi .

Shell shape. The most universally recognised character of A. nouryi is reportedly the elliptical shape of the shell: “The whorls increase in size very rapidly and the last is very elongate. Viewed laterally it is much shallower than is usual in the genus” ( Robson, 1932, p. 198). The shells are said to be “more elliptical than that of A. cornutus ” ( Keen, 1971, p. 895) and “longer than in any other species of Argonauta ” (Voss, 1971a, p. 33).

To investigate variation in shell shape across the lot, ShB was plotted against ShL (fig. 6). Probability plots indicate that both ShB and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, ShB the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in shell shape exists between eared and earless shells (F

(1, 136)

= 5.58, p = 0.02). Rib count. Argonauta nouryi shells are reported to have more ribs than A. cornutus shells: the ribs in A. nouryi are “more numerous” than in other species of Argonauta , while A. cornutus is reported to have “few radial ribs” (Voss, 1971, p. 32–33).

To investigate variation in the number of ribs per shell across the lot, RC was plotted against ShL (fig. 7). Probability plots indicate that both RC and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, RC the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in the number of ribs per shell does exist between eared and earless shells (F (1, 136) = 21.2, p <0.001).

Other features. To investigate the full range of quantifiable shell characters across the lot, scatter plots were similarly generated to investigate KTC, ApL, ApW and KW.

Keel tubercle count. To investigate variation in the number of keel tubercles per shell across the lot, KTC was plotted against ShL (fig. 8). Probability plots indicate that both KTC and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, KTC the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in the number of keel tubercles per shell does exist between eared and earless shells (F (1, 136) = 51.66, p <0.001).

Aperture length. To investigate variation in the length of the shell apertures across the lot, ApL was plotted against ShL (fig. 9). Probability plots indicate that both ApL and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, ApL the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in the length of the aperture does exist between eared and earless shells (F (1, 136) = 18.63, p <0.001).

Aperture width. To investigate variation in the width of the shell apertures across the lot, ApW was plotted against ShL (fig. 10). Probability plots indicate that both ApW and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, ApW the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in the width of the aperture does exist between eared and earless shells (F

(1, 136)

= 4.07, p = 0.046).

Keel width. To investigate variation in the width of the shell keels across the lot, KW was plotted against ShL (fig. 11). Probability plots indicate that both KW and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the linear regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, KW the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are equal and hence a significant difference in the width of the keel does not exist between eared and earless shells (F

(1, 136)

= 0.87, p = 0.353).

Statistical analysis indicates that significance differences in shell dimensions was associated with the presence or absence of ears. Eared shells have significantly lower RC (p <0.001), lower KTC (p <0.001), shorter ApL (p <0.001), increased ShB (i.e. shortened; p = 0.02) and increased ApW (p = 0.046). Earless shells have significantly higher RC (p <0.001), higher KTC (p <0.001), longer ApL (p <0.001), reduced ShB (i.e. elongate; p = 0.02) and reduced ApW (p = 0.046).KW was found to not be significantly different between shell types (p = 0.353).

Historically, the features of eared and earless shell types have been considered to represent separate species such that features of eared shells are considered characteristic of A. cornutus , while features of earless shells are considered characteristic of A. nouryi .

Two types of shell formation

Close examination of individual shells revealed that features considered characteristic of each shell type could occur on a single shell. While individual shells could display features of both eared and earless shell types, the characters did not appear in isolation. Sequential growth sections of the shells appeared to display all the characteristics of one shell type or another. For example, the initial component of the shell (the smallest whorl) may display all the characters historically associated with an A. cornutus shell while the latter component (the larger final whorl) may display all the features associated with an A. nouryi shell.

The most dramatic examples were shells that appeared to have been repaired over the course of the argonaut’s life. Fig. 12 presents photographs of one such shell from lot SBMNH 357476 (52.3 mm ShL). The initial component of the shell clearly displays the features historically attributed to A. nouryi (numerous fine ribs, reduced keel tubercles and no apparent ears), while the later component, following the clear repair line, displays a transition to features historically attributed to A. cornutus (ribs reduced in number and more pronounced, keel tubercles reduced in number and of larger size, and initiation of ears).

The presence of both shell types on a single shell clearly demonstrates that they represent different types of shell formation, not different argonaut species. This observation is supported by morphological evidence; despite full examination of nine female argonauts with shells (six historically identified as A. cornutus and three A. nouryi ), no morphological characters could be found to separate specimens with different shell types (see Finn, 2013).

The realisation that the two shell morphs represented two shell formation types, not two argonaut species, required that they be defined independent of previous species association:

• Type 1 shell formation (historically attributed to A. cornutus shells) – formation of ears, few pronounced ribs, few large keel tubercles, appearance of more pronounced arch in the keel resulting in a tighter final whorl (i.e. increased ShB, reduced ApL).

• Type 2 shell formation (historically attributed to A. nouryi shells) – absence of ears, numerous less pronounced ribs, numerous small keel tubercles, appearance of less pronounced arch in the keel resulting in the appearance of a shallower final whorl and elliptical shell (i.e. reduced ShB, increased ApL).

An important character associated with Type 2 shell formation is inter-keel tuberculation (tubercles on the keel surface; see fig. 2e). The appearance of inter-keel tuberculation on the keel of a shell flags a shift to Type 2 shell formation, while a loss of inter-keel tuberculation signifies a shift to Type 1 shell formation.

Based on this realisation, it became clear that this large lot, and all other material examined of these shell morphs, belonged to a single species. Because A. nouryi Lorois, 1852 , has date priority over A. cornutus Conrad, 1854 ; this study treats A. nouryi as the available name. See Finn (2013) for full synonymy. The key to understanding shell variation

The realisation that individual shells may be composed of combinations of two types of shell formation provided the key to understanding the huge variation in shell shape across the single large collection of argonaut shells from Baja California. Combinations of sequential shell formation could be recognised in all shells and hence their varied appearance could be understood. Shells were recognised within this single lot that display a single type of shell formation plus those with one, two or three transformations between the two shell formation types.

The initial whorl of most of the shells displayed Type 1 formation. Shell #37 displays a single change from Type 1 to Type 2 shell formation (fig. 13). Shell #72 displays a change from Type 1 to Type 2 shell formation and then a change back to Type 1 (fig. 14). Shell #41 displays a change from Type 1 to Type 2 shell formation and then a change back to Type 1 and then to Type 2 (fig. 15). Damage to shells normally results in a conversion to Type 2 shell formation.

In a transition between shell formation types, ears may be formed or subsumed. This is displayed across many shells within the lot. For examples, shell #139 displays subsumed ears as a result of a transition from Type 1 to Type 2 shell formation (fig. 16), while shell #136 displays ear formation, separate from the axis of the shell, as a result of a transition from Type 2 to Type 1 shell formation (fig. 17).

Type material. Available type material for additional species synonymised with A. nouryi Lorois, 1852 , was also examined for shifts in shell formation type. The holotype of A. dispar Conrad, 1854 (54.9 mm ShL, ANSP 129978) displays a single change from Type 2 to Type 1 shell formation (fig. 18). The holotype of A. expansus Dall, 1872 (80.2 mm ShL [P], USNM 61368), displays two changes – from Type 1 to Type 2 and then back to Type 1 (fig. 19).

Shell thickness. Preliminary observations suggested that the shell walls of Type 1 formation are thicker than the walls of Type 2 formation. To investigate this phenomenon, a scanning electron microscope was used to examine variation in shell thickness across recognisable shell breaks that corresponded with a switch between shell types (a single damaged shell from lot SBMNH 357476 was sacrificed). Preliminary results indicate a reduction in shell wall thickness between Type 1 and Type 2 formation. Fig. 20 presents two scanning electron micrographs displaying a reduction in thickness across a break signifying transition from Type 1 to Type 2. Shell thickness on the lateral face drops from approximately 220 to 140 µm (fig. 20a), while thickness at the keel drops from approximately 275 to 210 µm in this shell (fig. 20b) .

A lack of material that could be fragmented for examination with a scanning electron microscope limited the extent to which this phenomenon could be investigated. The lots housed in the SBMNH collection are too valuable to be considered for this style of destructive investigation.

A reduction in shell wall thickness may be related to producing a larger shell area with less shell material. The resulting thinner walled shell (Type 2) would therefore consist of less calcium carbonate and weigh less than an equivalently sized thicker walled shell (Type 1). The relative weights of the three shells presented in fig. 5 appear to support this theory. The Type 1 shell (cornutus-type voucher; 4.0 g) is 1.3 times the weight of the Type 1/ Type 2 shell (intermediate voucher; 3.1 g) and 1.7 times the weight of the Type 2 shell (nouryi-type voucher; 2.4 g), despite the shells having similar ShL. Weight (g) to length (mm) ratios of the three shells were: 1:16 for the Type 1 shell (cornutus-type voucher); 1:20 for the Type 1/ Type 2 shell (intermediate voucher); 1:28 for the Type 2 shell (nouryi-type voucher). These ratios suggest that per gram of calcium carbonate, Type 2 shell production results in a shell 1.8 times the length of a Type 1 shell.

To investigate this relationship across the lot, ShW was plotted against ShL with eared and earless shells distinguished (fig. 21). The scatter plot indicates a separation between eared and earless shells based on weight. This difference was analysed statistically to determine significance. Probability plots indicate that both ShW and ShL follow normal distributions. An ANCOVA was used to determine if the slopes of the regression lines, generated for eared and earless shells, were the same or different. Shell type (eared or earless) was the independent variable, ShW the dependent variable and ShL the covariate. The ANCOVA revealed that the slopes of the regression lines are not equal and hence a significant difference in weight exists between eared and earless shells (F

(1, 136)

= 86.7, p <0.001).

Figures

Figure 4. Comparison of a shell from the examined SBMNH lot with the type specimen and illustrations taken from the original description of Argonauta cornutus Conrad, 1854: a, reproduced illustration taken from the original description of A. cornutus Conrad, 1854, plate 34, fig. 2; b, photographs of the type specimen illustrated in the original description (58.6 mm shell length, ANSP 63496; please note, the original description illustrations mirror the characters of the shell, most likely due to the engraving and printing process of the era); c, illustrations of a shell matching the description of A. cornutus taken from the examined lot (shell #74, 65.0 mm shell length, SBMNH 345766). Illustration: R. Plant. Scale bar = 1 cm.

Figure 5. Three similarly sized shells of varied appearance selected from the examined SBMNH lot: a–c, three similarly sized shells of varied appearance selected from the single lot collected at El Mogote, La Paz, Baja California Sur, Mexico (24° 10' 00" N, 110° 24' 00" W) on 15 January 2000; a, shell #74 (65.0 mm shell length, SBMNH 345766) assigned the name cornutus-type voucher; b, shell #42 (61.2 mm shell length, SBMNH 345766) assigned the name intermediate voucher; c, shell #109 (66.5 mm shell length, SBMNH 345768) assigned the name nouryi-type voucher; i–iv, multiple perspectives of each shell; i, left lateral view; ii, anterior aperture view; iii, posterior keel view; iv, ventral view. Scale bar = 1 cm.

Figure 12. Repaired shell displaying components consistent with Argonauta nouryi and A. cornutus: a–d, four perspectives of a single shell (52.3 mm shell length, SBMNH 357476) displaying an initial component consistent with A. nouryi Lorois, 1854 (“nouryi”) followed by a subsequent component consistent with A. cornutus Conrad, 1854 (“cornutus”); a, right lateral view; b, oblique right lateral view; c, anterior aperture view; d, oblique ventral keel view. Dashed line represents repair line separating two visually different components. Scale bar = 1 cm.

Figure 20. Scanning electron microscope images of Argonauta nouryi shell displaying variation in shell thickness: a–b, scanning electron microscope images of shell cross-sections (SBMNH 357476) across shell repairs (R) representing a shift from Type 1 (T1) to Type 2 (T2) shell formation; a, lateral face of shell, inner surface facing up; b, keel, outer surface facing up. Scale bars = 1 mm.