Veins occur in many forms, are composed of many different minerals and occur at all levels of the Earth's crust and mantle. Their morphology, petrology and chemistry is a valuable source of information in a range of geological disciplines. The association of many ore deposits (particularly gold) with veins makes them even more relevant to geology. It is therefore not surprising that veins have been studied extensively. Yet, the formation of veins is still not fully understood. One surprising aspect of veins is that no one has been able to successfully and consistently simulate the formation of the variety of vein types - not in real rocks, nor in rock analogues. Attempts with varying degrees of success have certainly been made (Post 1989, Li & Means 1995, Means & Li 1995a&b, Bons & Jessell 1997, and probably many more which never have been published).
(Fig. 1)

Figure 1. Vein like textures in a thin sheet of octachloropropane (OCP, C3Cl8) between glass plates. OCP is an organic material, that has been used to simulate microstructural developments in crystal-plastically deforming rocks (Means 1989). Here, a sheet of fine-grained OCP was heated on a hot plate to its melting point (160°C) and immediately taken off when melting started, to prevent melting of the whole sample. Sliding of the top glass plate relative to the lower plate, caused extension between different parts of the partially molten sample, with immediate crystallisation out of melt occurring within the extension zones, resulting in vein-like structures. The whole process took place in a few seconds. (a) Whole "vein" with a curved stretch-crystal type texture. Width of view 5 mm, crossed polars. (b) Detail of a "vein" with an elongate blocky / stretched crystal texture and serrate ("radiator") grain boundaries, typical for stretched crystals, but possibly due to recrystallisation. Width of view 1.2 mm, crossed polars.

It is a pleasure to be able to state in this volume, that probably the most interesting and real-life-looking vein simulations were done under the supervision of Win Means (Li & Means 1995, Means & Li 1995a&b). However, these experiments and those by others, have not yet resulted in full understanding of the formation of veins. This paper will therefore focus on what is known about vein formation. First, the various types of veins and their (micro-) structures are discussed. Secondly, the modes of transport of vein forming material are reviewed, followed by the third and last section that deals with the processes and circumstances that lead to precipitation of vein forming minerals. The processes that lead to vein formation are linked with the structures that are found in veins. It is impossible to go into detail of each individual aspect of vein formation within the limited space of this paper. Some topics are therefore only dealt with briefly, especially when much literature on the topic is available (e.g. veins and fluid flow through fractures). Other topics are discussed in more detail, especially if relatively little is published on these topics, such as the two end members of vein forming processes: vein growth without fracturing and veins formed out of mobile hydrofractures.

Before proceeding to discuss the formation of veins, it is important to define veins. In this paper, I define veins as "distinct polycrystalline mineral volumes that formed within a rock and that are filled with one or more minerals that precipitated from an aqueous fluid". The term "polycrystalline" figures in this definition to exclude individual metamorphic porphyroblasts. "Formed within a rock" is added to exclude evaporitic precipitates, although such precipitates can also form within sediments. Definitions of veins often include a description of the shape of veins, typically planar or lenticular. I do not do so, as veins have many shapes and I specifically want to classify pressure fringes as veins. Finally, small igneous bodies are also often termed "veins", but these are excluded in the definition given here as their melt origin is different from that of the veins discussed in this paper. 

One could say that veins are structures that reveal a deformation history for structural geologists (e.g. Ramsay & Huber 1983), while veins are principally indicators of past fluid flow for metamorphic petrologists and geochemists (e.g. Thompson 1997). This difference in approach has not helped the issue. We will only achieve real understanding of veins if we combine all of these aspects and also if we consider the full range of vein-forming processes and the resulting vein types. In this paper I review the terminology of veins (traditionally mostly the domain of structural geologists) and the processes that lead to vein formation (more the field of metamorphic petrologists), with the aim of clarifying how different vein types form and how to recognise the processes that led to vein formation.