This paper identifies and describes the textual densities of ideational metaphors through the application of GM theory (Halliday, 1994) to the textual analysis of two twentieth century English short stories: one American (The Mansion (1910-11), by Henry Jackson van Dyke Jr.), and one British (Home (1951), by William Somerset Maugham). One aim is to get at textually verifiable statistical evidence that attests to the observed dominance of GM nominalization in academic and scientific texts, rather than to fiction (e.g. Halliday and Martin (1993). Another aim is to explore any significant differentiation in GM’s us by the two short- story writers. The research has been carried out by identifying, describing, and statistically analysing the frequencies of ideational GM structures in both fiction texts to get at their comparative textual densities in terms of word-counts. The obtained results have shown that GM structures – though used in both the American and British short stories – are statistically quite infrequent in both texts, accounting for a tiny (0.0064%) of the total text-wording in T1. against (0.0137%) for T2. Such very low rates of frequency (well below the threshold of even 1% of each text volume) corroborates the previously observed dominance of GM nominalization in academic and scientific texts, rather than in fiction. These same low densities of use does not allow drawing significant inference differentials in GM’s use by the two writers.