H.M. is a 17-year-old girl of at least average intelligence and with an above-average oral vocabulary. She is impaired at non-word reading in comparison to word reading. She is incapable of reading any long nonsense words or long unusual regular words correctly. Many of her non-word reading responses are lexicalizations. She reads non-words homophonic with real words better than she reads non-words that are not homophonic with real words. Her reading age is 10 years 11 months. A large proportion of her reading errors are derivational or visual paralexias. She makes no semantic errors or errors of regularization and few neologistic responses. H.M. is not influenced by the dimensions of spelling-to-sound regularity and word length. There is no effect of word class on single-word reading although a function word deficit is present when continuous text is read aloud. H.M. is impaired at word reading when the stimulus items are distorted in a manner that reduces the potential for global perception. H.M.'s spelling errors are primarily phonological, though more complex errors are made than those characteristic of surface agraphics. All of these features are consistent with reported cases of acquired phonological dyslexia. H.M. may be confidently regarded as a developmental phonological dyslexic.

One form of developmental difficulty with arithmetic affects the storage or retrieval of arithmetical facts, such as tables, which are required to implement arithmetical computations (Temple, 1991, 1994). Such difficulties may arise because of impairment in a specialized system for the storage of arithmetical facts or as a result of causally linked impairment in another cognitive domain. This study explored issues concerning the representation and retrieval of arithmetical facts in children with number fact disorders (NF) and in normal children, in particular the status of hypothesized linked impairments: short-term memory (STM) spans, counting skills, speed of speech, and speed of number fact and lexical retrieval. There was no evidence that NF children had weak STM spans on any span measure or that STM spans related to arithmetical fact skills. There was also no evidence that NF children had weak counting abilities or free counting speeds. The NF children were slower in speeded counting, which also correlated with number fact skill. The significance or not of this is discussed. The NF children were also slower than controls in speed of speech and on some measures of speed of access. However, the absence of correlation with number fact skill, the absence of generality across tasks, and the possibility that delayed speeds in fact retrieval reflect the use of alternative strategies, together suggest that the increased speeds are not causally linked to number fact skill. The results are consistent with modular accounts, in which there is a specialized system for the storage and retrieval of arithmetical facts.

Traditional developmental models describe a series of sequenced stages and interpret developmental disorders in teim of arrested development. Two different cases of developmental dyscalcullia are described which cannot both be explained against the same stage model of the development of axithmetical skill. In both cases accurate number processing skills have developed.SW has also acquired accurate number facts and tables but has selective difficulty with the procedures of calculation. HM shows the reverse pattern with normal mastery of calculation pr0cc:dures but selective dif6culty in the mastery of arithmetical tables. The arithmetical modules involved in tables and procedures develop in a semi-independent fashion, neither being an essential precursor of the other. This study demonstrates that there are individuil differences in the developmental dyscalculias consistent with individual differences in the developmental pathways to the adult calculation system. It alsodemonstrates that the developmental dyscalculias are analapus to the acquired dyscaldias.

No abstract