The term allotrope is one of the most vaguely defined chemical terms still in current use. I can understand how one could call the various forms of sulfur (ie. rhombic and monoclinic) as distinct allotropes. But 2 chemical species that can't be transformed into each other by physical means just don't seem to garner the right to be called an allotrope. This issue was recently brought up by Fritnat, http://www.chemicalforums.com/index.php?topic=8500.0, when he advocated ozone should be called oxygen since it is an "allotrope" of diatomic oxygen.
But, are we to say that anything with the same empirical formula is an allotrope? I doubt I would find a single chemist who would argue octane and methane are both allotropes since they have the same empirical formula CnH2n+2.
At any rate you can read more about the wonderful world of allotropism in a recent article by William B. Jensen, http://jchemed.chem.wisc.edu/Journal/Issues/2006/Jun/abs838.html. It's a very short read and packed with a historical perspective on the use of the word allotrope. Eventhough, Mr. Jensen never replied to a comment I sent him, I would still recommend everyone read this article.
Mitch
Many words change its meaning during time beeing. But, I think, allotrope has practically constant meaning from early decades of XX century. Its means - different chemical species of the same element - hence diamond, graphite and fullerenes are allotropes, but diamond and lonsdaleite, or hexagonal graphite and rhombohedral graphite are pairs of polymorphs.
For me, polymorphism means different crystallographic symmetry of the same chemical species (compounds or elements), and isomerism means different arrangement of atoms in the same molecular formula.