Multivariate analysis is a statistical method used to study the joint relationships of variables in data with intercorrelations. It allows for simultaneous consideration of multiple variables, enabling interpretations not possible with univariate statistics. The first applications of multivariate analysis in ecology and systematics were in plant ecology and numerical taxonomy over thirty years ago. Recent literature shows widespread use of multivariate methods in ecology, ordination, classification, wildlife biology, systematics, and morphometrics. Between 1978 and 1988, numerous books, symposia proceedings, and reviews were published on these topics. From 1983 to 1989, 514 applications were found in seven journals. Multivariate analysis is essential for understanding ecology and systematics. Misunderstanding the methods can hinder scientific progress. The paper discusses three controversial topics: multicollinearity, indirect ordinations in plant ecology, and the study of shape variables in morphometrics. It argues that shape variables, defined as ratios and proportions, should be studied directly. The paper also notes that multivariate methods can be misused, leading to circular reasoning and incorrect conclusions about causation. The paper summarizes various multivariate methods, their objectives, limitations, and examples. It emphasizes that linear methods are appropriate when variables are continuous and linear combinations are needed, while nonlinear methods are more suitable for categorical or ranked data. The paper warns against the misuse of stepwise procedures and highlights the importance of proper statistical inference. It also discusses the importance of biological knowledge in interpreting multivariate results. The paper concludes that multivariate analysis is a powerful tool in ecology and systematics, but its proper application requires careful consideration of the data and the research objectives. It emphasizes the need for a balanced approach that combines statistical methods with biological understanding to avoid misinterpretation and ensure accurate conclusions.Multivariate analysis is a statistical method used to study the joint relationships of variables in data with intercorrelations. It allows for simultaneous consideration of multiple variables, enabling interpretations not possible with univariate statistics. The first applications of multivariate analysis in ecology and systematics were in plant ecology and numerical taxonomy over thirty years ago. Recent literature shows widespread use of multivariate methods in ecology, ordination, classification, wildlife biology, systematics, and morphometrics. Between 1978 and 1988, numerous books, symposia proceedings, and reviews were published on these topics. From 1983 to 1989, 514 applications were found in seven journals. Multivariate analysis is essential for understanding ecology and systematics. Misunderstanding the methods can hinder scientific progress. The paper discusses three controversial topics: multicollinearity, indirect ordinations in plant ecology, and the study of shape variables in morphometrics. It argues that shape variables, defined as ratios and proportions, should be studied directly. The paper also notes that multivariate methods can be misused, leading to circular reasoning and incorrect conclusions about causation. The paper summarizes various multivariate methods, their objectives, limitations, and examples. It emphasizes that linear methods are appropriate when variables are continuous and linear combinations are needed, while nonlinear methods are more suitable for categorical or ranked data. The paper warns against the misuse of stepwise procedures and highlights the importance of proper statistical inference. It also discusses the importance of biological knowledge in interpreting multivariate results. The paper concludes that multivariate analysis is a powerful tool in ecology and systematics, but its proper application requires careful consideration of the data and the research objectives. It emphasizes the need for a balanced approach that combines statistical methods with biological understanding to avoid misinterpretation and ensure accurate conclusions.