Morphological changes observed in ammonites at the Lias-Dogger (Toarcian-Aalenian) boundary are analyzed here. The study covers all known ammonites worldwide over some 2 m.y. (the last two biozones of the Toarcian: Pseudoradiosa and Aalensis and the first biozone of the Aalenian: Opalinum). Ammonite taxonomy at this boundary is still open to question, the main disagreements being over groupings at family or super-family level. The evolutionary schema developed by Tintant and Mouterde  shows two main sets of Ammonitina for the time: (1) the dominant Hammatocerataceae (Graphoceratidae+Hammatoceratidae) and (2) the final representatives of the Hildocerataceae. Two conservative ammonite stocks coexisted alongside these Ammonitina: Phylloceratina and Lytoceratina. The ammonites under review are those that made the transition from the "typically Liassic stock" to the "post-Liassic stock" prefiguring all subsequent Jurassic ammonites. A morphometric procedure is used to quantify ammonite morphological disparity rather than taxonomic diversity. A set of 35 characters were considered, corresponding to the end of the phragmocone and to the body chamber. We selected 64 morphologies covering the entire range of shapes found at the subzone scale. Morphological disparity is expressed graphically by morphospaces (first factorial plane of correspondence analysis) and arithmetically by a space occupation index (sum of variances on the factor axes). The FAD and LAD of the selected species allow us to study morphospace changes at the subzone scale. The results clearly show changes occurring in the course of the seven subzones. The earliest subzone (Levesquei) has points weakly grouped around two poles (negative and positive F1 values) with an isolated point characterized by a very negative F2 value. Changes in this morphospace from the Pseudoradiosa subzone to the Lugdunensis subzone are marked by increased dispersion around the two poles. By contrast, an increasing number of morphologies have clearly negative F2 values. The Lias-Dogger boundary displays further dispersion of the two morphological poles, leading to them becoming indistinct. Continuation after the boundary shows a very similar pattern with shapes scattered and covering most of the morphospace. The morphospace changes analyzed previously were quantified by the sum of variances of the first 15 factorial axes, representing more than 90% of the information contained in the original matrix. For the Ammonoidea as a whole, the poorest coverage of the morphospace occurred in the Levesquei subzone, while the best coverage occurred in the Comptum and Bifidatum subzones. Between-times, morphological disparity increased at varying rates with no drastic changes at the Lias-Dogger boundary. A similar procedure was performed for the Ammonitina only. Indeed, the occurrence of two conservative sets (Lytoceratina and Phylloceratina) could be suspected of inducing -- at least initially -- two peculiar morphological poles and introducing a bias into the quantification. The results are similar, bringing out more clearly the increase in morphological disparity at the end of the Lias, and its stagnation thereafter. It is difficult to say why morphological disparity increased as it did. The main problem is the lack of any clearly resolved phylogeny for the ammonites under study. However, if we accept the hypothesis of Tintant and Mouterde  recognizing five homogeneous clusters for ammonites of these times, it can be seen that two of them alone (Graphoceratidae first, and Hammatoceratidae later) account for the increased disparity. Within this phylogenetic assumption, it is improbable that external constraints alone, such as eustasy or physico-chemical phenomena, could have brought about the increase, for they would probably have affected all of the ammonoids, or at least the three Ammonitina groups that were initially quite similar in shape (in the Levesquei sub-zone). It is more likely then that internal factors (e.g. heterochrony) leading to two consecutives biological "explosions" (Graphoceratidae and Hammatoceratidae) were instrumental in bringing about this increased disparity. Finally, if the Lias-Dogger boundary is clearly marked by changes at any taxonomic level, our results--based on a quantification of morphological disparity and not on taxonomy--show that changes in ammonites can only be properly understood in a broader context: there was some morphological change at the Lias-Dogger boundary but it was initiated at the beginning of the Pseudoradiosa subzone. The Lias-Dogger event was thus not an unusual or major one, but more the end of a sustained event initiated two zones earlier. This period of time (final two zones of the Toarcian) witnessed progressive growth in disparity, which ceased only at the beginning of the Dogger.