This review article adresses a number of issues in the reconstruction of Proto-Indo-European inflectional and derivational nominal morphology, raised in a recent excellent book by P. Widmer. While this book contains many important insights and astute observations, some of its claims are harder to accept. In particular, contrary to Widmer’s claim, there is no reliable evidence for Schwebeablaut in internal derivation: the reconstruction *h2u̯és-es from which familiar *h2éu̯s ōs ‘dawn’ should have been derived, lacks conviction, since W.’s two pieces of evidence (*h2u̯és-er > Vedic vasar(-hán-) and *h2u̯és-en > Tocharian А wṣe, В yṣiye) are both better explained as delocatival hypostases. Same is true for two other examples, *h3enbh-en- and *g̑héi̯m-en-. Instead, the review article argues that Schwebeablaut, descriptively a marker of PIE locatives with suffixes *-er /*-en /*-el, is in fact due to vrddhi in the root: the morphological operation that inserted an extra full-grade vowel in the root (preferably at a “wrong” place) goes hand in hand with the derivation of locative caseforms in PIE and should be viewed together with lengthened grades in such locatives as *dēm or *pēd. Several other details of morphological reconstruction are critically discussed, such as the accentual difference between Vedic vrṣṇi-, n. and vrṣṇí , m. that purportedly reflects different prehistorical accents of neuter and animate.