Since its donation to TMAG by the Gellibrand family in 2004, this portrait has presented a puzzle. Unsigned and undated, it was originally thought to be by the colonial painter, Benjamin Duterrau. Later, due to stylistic comparisons, the painting was attributed to Augustus Earle – an itinerant, yet highly regarded, artist who briefly resided in Van Diemen’s Land in 1825.

However, the attribution to Earle is problematic. Firstly, at the time that Earle was living on the island, the Gellibrand family had only three of their eventual nine children – and the portrait shows four. Secondly, when the fourth and fifth children were born, Earle was no longer living in the state. Thirdly, although Earle did return briefly to Van Diemen’s Land in 1828, by this time the Gellibrand’s had five children. In order to match the biographical ages and identities of the children, the best date for the painting would be 1827 – in which case, it would show Thomas (aged seven), Eliza (five or six), William (three or four) and Joseph (12–18 months) – but Earle was living in New South Wales at the time.

Portraits of children in the early 19th century were uncommon - few artists were working in the colonies in the 1820s and the child population was low. Some scholars have suggested that the painting is made from sketches that Joseph Gellibrand (1792–1837), Tasmania’s first attorney general, may have taken to Sydney with him when he was admitted as a barrister of the Supreme Court of New South Wales in 1827. It is possible that he visited Earle on this occasion and commissioned the portrait. Certainly, the ‘levitating’ position of the baby in the front, the awkward entwined arms of the youngest children and the finger twirling the necklace on the second oldest child all strongly suggest it was a painting composed from different sketches.

Regardless of the composition, the mystery of the missing artist and date continues to intrigue today.