Ontogenetic shifts in natural diet during benthic stages of American lobster (Homarus americanus), off the Magdalen Islands

The natural diet of 506 American lobsters (Homarus americanus) ranging from instar V (4 mm cephalothorax length, CL) to the adult stage (112 mm CL) was determined by stomach content analysis for a site in the Magdalen Islands, Gulf of St. Lawrence, eastern Canada. Cluster and factor analyses determined four size groupings of lobsters based on their diet: <7.5 mm, 7.5 to <22.5 mm, 22.5 to <62.5 mm, and ≥62.5 mm CL. The ontogenetic shift in diet with increasing size of lobsters was especially apparent for the three dominant food items: the contribution of bivalves and animal tissue (flesh) to volume of stomach contents decreased from the smallest lobsters (28% and 39%, respectively) to the largest lobsters (2% and 11%, respectively), whereas the reverse trend was seen for rock crab Cancer irroratus (7% in smallest lobsters to 53% in largest lobsters). Large lobsters also ate larger rock crabs than did small lobsters.

On the identification of the early developmental stages of clupeiform fishes from Indian waters

Among about 65 species of clupeiform fishes present in the seas around India, most are distributed along both the east and west coasts and have similar spawning seasons and spawning grounds. The most difficult problem experienced with regard to the identification of their early developmental stages is the overlapping sets of diagnostic features among many species within the same genus as well as among species belonging to two or more genera. From studies carried out recently, the present paper sites a few instances wherein certain subtle characters have been used for distinguishing the early developmental stages of a few allied species.

On the unknown postlarval stages of the garfish, Tylosurus strongylurus (Van Hasselt) and the seapike, Sphyraena jello Cuvier

Unknown larval stages of Tylosurus strongylurus and Sphyraena jello have been traced from Vellar Estuary, Porto Novo (lat. 11 degree 29'N; long. 79 degree 46' E). Descriptions of different stages (7-17.5 mm T. strongylurus, and 18 and 22.5 mm S. jello) are given. A thorough ichthyoplankton survey for 2 years (Nov. 1977-Oct. 1979) in the estuary revealed that these larvae were very rare in the estuary and were caught on the occasion (Nov. 1978) only.

Laboratory reared eggs and larval stages of the clupeid fish, Opisthopterus tardoore (Cuvier)

The early embryonic and larval stages of Opisthopterus tardoore are described and discussed. The developing 3 eggs of stages up to 21hrs after collection and larvae up to 118hrs after hatching were reared in the laboratory. Seasonal occurrence of the eggs in the Vellar estuary, Porto Novo (lat. 11 29'N; 79 degree 46'E) was recorded for a period of 2 years (Nov. 1977 to Oct. 1979). Variation in myotomic counts in the larvae is critically reviewed with earlier works.

Operatives training programmes in India

During the last few years great changes have taken place in the fishing industry as a result of which production of fish in the world has increased enormously. From an insignificant trade employing tools and methods of primitive nature, fishing in many countries has become an important industry utilizing complex modern vessels equipped with electronic equipment and operating in the high seas with highly mechanized fishing gear.

Competitive training in hidden Markov models

The use of hidden Markov models is placed in a connectionist framework, and an alternative approach to improving their ability to discriminate between classes is described. Using a network style of training, a measure of discrimination based on the a posteriori probability of state occupation is proposed, and the theory for its optimization using error back-propagation and gradient ascent is presented. The method is shown to be numerically well behaved, and results are presented which demonstrate that when using a simple threshold test on the probability of state occupation, the proposed optimization scheme leads to improved recognition performance.

MMI training for continuous phoneme recognition on the TIMIT database

This paper reports our experiences with a phoneme recognition system for the TIMIT database which uses multiple mixture continuous density monophone HMMs trained using MMI. A comprehensive set of results are presented comparing the ML and MMI training criteria for both diagonal and full covariance models. These results using simple monophone HMMs show clear performance gains achieved by MMI training, and are comparable to the best reported by others including those which use context-dependent models. In addition, the paper discusses a number of performance and implementation issues which are crucial to successful MMI training.