Manganese and its inorganic compounds are widely used in many industries and have been accepted as occupational neurotoxins that have caused a distinct and disabling clinical entity, manganism, in several types of work, notably where exposure is by way of dust. There is inconclusive and inconsistent evidence that, in these occupations, subclinical neurological effects, detectable only by neurobehavioural studies, may be caused by low doses. This has prompted a re-evaluation of occupational exposure limits. Some countries, including the UK, already demand much higher levels of protection against exposure than 5 years ago. Welding is the most common source of occupational exposure as manganese is an essential component of steel and so its compounds are inevitable components of fume emitted from steel welding processes. There it is found in respirable particles, often as complex oxides (spinels), sometimes within a core protected by a silicon oxide shell - as distinct from the much simpler form of particle formed by disintegration in processes such as mining and ore milling where manganism has been diagnosed convincingly. Millions of workers are at risk of exposure to manganese-containing compounds in fumes from electric arc welding of steel. In recent years it has been asserted that neurological and neurobehavioural disorders may develop consequent to exposure to steel welding fumes and that employment as a welder is associated with the unusually early onset of Parkinson's disease. Causal relationships have been postulated. Welders have been recorded as having been exposed to high levels of manganese-containing fume, especially where they have worked in confined, unventilated spaces, although this appears from limited data to be the exception rather than the rule. Even then the dose received is generally less than in mining or ore crushing. When care is taken to exclude exposures from hardfacing and burning and cutting arc processes, where manganese may form a high percentage of the fume, manganese compounds usually form a relatively low percentage of the composition of welding fume particles, <2.0%, much outweighed by iron. Although these manganese-compound-containing welding fume particles are insoluble in water, the manganese compounds in particles that are retained in the alveoli may be absorbed, at least in part. Manganese concentrations in biological material samples in some exposed groups reflect this relative to unexposed workers. Some of the transfer systems for absorption and transport, including across the blood-brain barrier, are used in competition with iron which is present in abundance in welding fume. This may reduce absorption of manganese in welders and thus reduce the opportunity for sufficient doses to cause neurotoxicological consequences. Scrutiny of the literature covering the last 40 years has revealed only five cases that meet sufficient criteria for manganism to just cross the diagnostic threshold, and even then they carry a degree of doubt with them. This low incidence alone gives notice that welders have not been and are not at high risk of clinically apparent damage from exposure to manganese. If this needs to be further emphasised, there is the fact that the literature contains no confirmed cases of manganism in welders. Assertions of abnormal results in neurobehavioural studies of welders have raised the possibility of there being a subclinical form of manganism with loss of fine motor control as one of its features. While observations of such changes in workers in other industries have caused regulators in some countries to apply more stringent controls of exposure, as yet the results lack convincing consistency and there is no indication of any dose-effect relationship. If welding fume can have these motor effects it would be a heavy and perhaps career-ending blow to those affected. It would not be prudent to dismiss the warnings sounded by the results of studies of welders, no matter how flawed these investigations are, but wiser and better to act with vigour to reduce exposure and monitor the effectiveness of this additional protection whilst conducting high quality research to allow sound conclusions to be drawn as to whether there actually is a subclinical disorder. Idiopathic Parkinson's disease is a common disorder affecting 1-2% of those in the general population aged >65 years. It has been suggested, on flawed and contested evidence, not that welding causes the disease but rather that employment as a welder carries with it the risk of developing this disease at a younger age than if that trade had not been followed. Manganese in welding fume has been nominated as the neurotoxin. This may be biologically feasible if manganese destroys insufficient receptor cells to produce clinical manganism but sufficient to enhance the effects of a reduced supply of dopamine to give the manifestations of already developing idiopathic Parkinson's disease earlier in the course of destruction of the substantia nigra than if all receptors were intact.