The paper focuses on the probability of the Celtic substratum hypothesis in the toponymy of North Holland. Agreeing that the most north-western tip of the Netherlands is an unlikely place to look for Celtic toponyms, the author suggests that the name Huisduinen relates to the same group of names of which Heusden is the most common representative, and which appears to have a Celtic etymology. Thus making it a tempting task to look at a few other names in the same area. As the area lost most of its population in the 4th century AD and became repopulated in the 5th century, language shift offers a possible scenario for a change from Celtic to Germanic with remnants of a Celtic substratum surviving up to the present day. In the same period, the landscape involved saw radical changes as well. In earlier publications it has been suggested that the medieval name Uxalia may be Celtic. Here it is suggested that this name may originally refer to the present-day island of Texel and not — as it later did — to the neighbouring island of Vlieland. A Celtic etymology is also proposed for the names Helsdeur and Den Helder, which — if accepted — have related etymologies. The name Helsdeur refers to the deepest part of the strait between the mainland of the province North Holland and the island of Texel. The lack of early attestations of this name is explained by suggesting its probable taboo status. This hypothesis is supported by a series of relevant examples of taboo place names in the maritime context.