The classification of the subclass Protobranchia followed here considers the families Nuculidae, Nuculanidae, Siliculidae, Sareptidae, Malletiidae, and Tindariidae in the order Nuculoida, and the family Acharacidae in the order Solemyoida. We also follow the separatation of Malletiidae from Nuculanidae, further justifying it on anatomical grounds. Fifteen protobranch species of Recent distribution in continental Chile, five Antarctic, and seven fossil species were studied. Out of a total of 35 species cited for the coast of Chile and Graham Land in Antarctica, only the following 27 are here accepted as valid: Nucula austrobenthalis Dell, 1990; Nucula (Nucula) falklandica Preston, 1912; Nucula (Nucula) fernandensis Villarroel, 1971; Nucula (Nucula) interflucta Marincovich, 1973; Nucula (Nucula) pisum Sowerby I, 1833; Ennucula eltanini Dell, 1990; Ennucula grayi (d'Orbigny, 1846); Ennucula puelcha (d'Orbigny, 1842); Nuculana (Saccella) cuneata (Sowerby I, 1833); Nuculana (Borissia) inaequisculpta (Lamy, 1906); Propeleda longicaudafaThiele, 1912; Tindariopsis sulcuta (Gould, 1852); Silicula patagonica Dall, 1908; Silicula rouchi Lamy, 1911; Yoldia (Aequiyoldia) eightsi (Jay, 1839); Yoldiella chilenica (Dall, 1908); Yoldiella ecaudata (Pelseneer, 1903); Yoldiella granula (Dall, 1908); Yoldiella indolens (Dall, 1908); Malletia chilensis des Moulins, 1832; Malletia magellanica Smith, 1875; Malletia patagonica Mabille & Rochebrune, 1889; Malletia inaequalis Dall, 1908; Malletiella sorror Soot Ryen, 1959; Tindaria virens (Dall, 1890); Tindaria salaria Dall, 1908; and Acharax macrodactyla (Mabille & Rochebrune, 1889). Nucula (N.) pseudoexigua Villarroel & Stuardo, a new species from the Strait of Magellan is described. The Antarctic species Propeleda longicaudata Thiele, 1912, is reported for the first time from the Strait of Magellan. A diagnosis for every taxonomic category and a detailed description for every species are given, including the shell features traditionally utilized, as well as the soft parts. Most features having so far been studied only on a few species of the subclass, allow a comparative analysis with the Chilean representatives, summarized as follows. (1) Size varies within the different families. Living nuculids are in general smaller than nuculanaceans and solemyaceans. The largest size found in the Chilean Nuculidae reaches 20.6 mm length in Ennucula grayi, whereas among Nuculanidae and Malletiidae, a maximum measured length of 51.0 mm was found in Malletia chilensis. A Chilean fossil of this genus measured 60 mm. (2) The studied species fall within the three known basic forms: nuculoid, nuculanoid, and solemyoid (Fig. 61). (3) No comprehensive study of hinge tendencies within each family has been attempted; such study would possibly permit to examine affinities and divergencies at lower taxonomic ranks. (4) The study of the ligament in specific taxa should be used to test the validity of prevailing models and interpretations. So far, the study of the ligament in Nuculidae and Nuculanidae has followed Owen's (1959) interpretation of an external or lamellar layer connected with the mantle margins, and another internal, fibrous layer connected to the isthmus of the mantle. A resilifer or chondrophore interrupts the two teeth series, and is directed anteriorly in Nucula and Ennucula (Fig. 3A, cdr), is more or less straight in Yoldia (Fig. 129), and is directed posteriorly in Nuculana (Fig. 3). Previously, Stempell (1898a) demonstrated that the ligament in Malletia can be divided in anterior, central and posterior parts, the central part corresponding to the resilium (inner layer of the ligament), and the anterior and posterior parts with an external origin. Thus, such similar differentiation in Nuculidae, Nuculanidae and Malletiidae, suggested that the resilium of internal position in Nucula and Nuculana, had migrated to become external in Malletiidae, without dissapearing. An intermediate stage in its position is observed in Tindariopsis, as was shown by Stempell (1898a). Relevance is given to the novel and most stimulating interpretation on the evolution of the ligament in the bivalves advanced by Waller (1990). He questions the traditional model of an amphidetic primary ligament of three layers, and proposes a protobranch stem group from which two major types of ligament for the bivalves evolved. (5) Variation in number and size of hinge teeth does not allow to use them as taxonomic features of generic or suprageneric value, but size and form may sometimes offer specific taxonomic value, as noted by Knudsen (1970) and Villarroel (1971). (6) The palps are very similar in the Nuculacea and Nuculanacea, but their homology with the Solemyidae is not well known. The palps in Solemya are not interpreted as doubled palps appendixes of other protobranchs; the palps sheets would be reduced to simple ridges (Fig. 11) in the edge of the furrow that joins the mouth with the appendixes (Figs. 15-17)(Ridewood, 1903; Morse, 1913; Yonge, 1939; Reid, 1980). The appendix or palp tentacle on the external sheet of every palp considered by Drew (1901) be equivalent of a pair of hypertrophiated fold (Figs. 10, 12, other figs., tp) differ in position according to family (Fig. 61). In the Nuculanacea, the palp appendix is located on the terminal portion of the external palp sheet (e.g., Fig. 5, Silicula rouchi; Fig. 77, Nuculana (S.) cuneata; Fig. 90, Nuculana (B.) inaequisculpta). In the Nuculidae, the palp appendix is displaced to the end, because behind it there is an additional, non-extensible structure termed the "palp caecum," which represents a pair of hypertrophiated folds (Stasek, 1965). The proximal end of the palp appendix is linked with the surface of the palp external sheets and the palp caecum (Fig. 10, bp); its musculature is fused with the posterior foot retractor. Stasek's (1961) observation of this feature in Acila was corroborated, without exception, in every species studied. Nevertheless, in almost all the cases, the appendix was found in different degrees of contraction, preventing recognition of specific differences (Figs. 4 and 62, 74 and 76, tp). (7) In a general comparative analysis, the value of the soft parts in the differentiation of the higher categories within the subclass, is corroborated; however, due to the limited available knowledge of many internal structures and the few studied species, their taxonomic role at the specific level cannot be always ascertained. On the other hand, the complexity observed in some of the internal morphological parts permitted us to set forth complementary interpretations on their possible phylogenetic value, particularly in the case of the stomach, the position of the heart, and the configuration of the various types of siphons. Although stomach morphology has been described for species of Nucula, Nuculana and Malletia, a comparison became necessary, resulting in the identification of a new caecum and changes in the interpretation of the features observed by previous authors. In fact, its study in the available species allowed the conclusion that there is not one basic type or "Gastroproteia," as proposed by Purchon (1956, 1959), but three. These are: Type Ia. Common to the genera Nucula and Ennucula and characterized by several (three or four) ciliary sorting areas and a wide extension of the typhlosole (Figs. 18-32, 60). Type Ib. Common to the genera of Nuculanidae and Malletiidae and characterized by three ciliary sorting areas and a small extension of the minor typhlosole (Figs. 33-56, 60). Type Ic. Common to the genera of Solemyidae and Nucinellidae and characterized by the absence of distinct sorting areas and lack of typhlosoles. It is not difficult to differentiate the internal and external features recognized in the stomach of Nuculacea and Nuculanacea. The dorsal hood is smaller in Nuculanacea than in Nuculacea, and the three ducts that communicate the stomach with the digestive diverticula are also different in these two superfamilies (Figs. 18-56, 60). Similar differences were found in the ciliary sorting areas and the number of folds. For instance, the three additional sorting areas as1, as2 and as3 described by Purchon (1956), although not present in all species, can also be used in interspecific differentiation. Thus, the first one was found only in Nucula (Nucula) pisum (Fig. 21) and Ennucula puelcha (Figs. 29-31), but not in the other studied species of these genera; the second sorting area was found presenting different sizes in Nucula (Nucula) pisum, Nucula (Nucula) fernandensis, and Ennucula puelcha, being largest in the latter. The third above named sorting area was not observed in the studied Nuculacea, and none were found in the studied Nuculanacea. Such differences do not back Purchon's (1987b) generalization that one description can embody all of them (Fig. 60). Undoubtedly, the complexity of the gastric shield with its biggest modification in Propeleda and Malletia is larger in Nuculanacea than in Nuculidae and Solemyidae, but presently it is difficult to establish generic or specific differences. On the other hand, folding of the typhlosoles entering the style-sac has shown specific constancy in the studied species of Nucula and Ennucula. Development of the typhlosoles in Nuculanacea shows a different pattern. (8) Attention has also been given to the number of loops observed in the gastric and medium intestine with a pattern of coiling, which according to Heath (1937) is specific, with minimal intraspecific variation as observed in Nucula (Nucula) pisum and Nucula (Nucula) pseudoexigua (Figs. 63-67). It begins on the side of the stomach and continues anteriorly in some species almost reaching the mouth. It turns then dorsally to the esophagus and continues posteriorly above the stomach, or continues ventrally to form the coils prior to its final turn backwards.