The development of a new grammar of English for machine translation systems at Weidner Communications Corporation is described. Although the project began with the intention of simply modifying the grammar rules already in use in the translation system, the reorganization of the grammar was deemed necessary for providing a principled manner of increasing coverage and for improving the efficiency of the grammar read by the parser. The most innovative aspect of the grammar is the approach to the internal structure of phrasal constituents. This approach, referred to as Left-Modifier Grouping, groups left-hand modifiers of most lexical heads with the head as a constituent to which right-hand modifiers are then attached. Left Modifier Grouping is based on a linguistic generalization about the phrase structure of English by which nodes to the left of all lexical heads except "V" are not themselves right-branching. A consequence of Left Modifier Grouping is that a practical grammar of English for parsing can be broken down into two types of rules: (1) those that are not recursive, thus applicable only once; and (2) those that are recursive, allowing reapplication. A summary of the two blocks of rules is presented. (DJD)