We welcome Barber’s (1) comments and are grateful for the opportunity to respond. Our study of wetland taro (Colocasia esculenta) gardens during the initial colonization period (ICP) (1200 to 1500 CE) in New Zealand did not overlook the evidence from the Aupouri Peninsula (2⇓–4). We agree that gardens were probably established on mainland New Zealand, within the climate envelope shown in figure 1 of our paper (4), but in areas that lacked large-statured forests at Polynesian arrival. However, the fossil evidence from Motutangi does not meet the 3 criteria for defining ICP taro gardens met in our study of Ahuahu and subtropical French Polynesia:

• 1) Reliable fossil proxies: We identify pollen, the most reliable fossil proxy for taro (5), but also clusters of small globular orbicular starch grains inside possible parenchyma cells, similar to those described by Horrocks and Barber (2). We doubt whether these can be used to distinguish taro, at least when using light microscopy to observe starches from indigenous New Zealand species. Barber (1) in his letter does not refer to the calcium oxalate raphides found at Motutangi, originally attributed to taro (2), perhaps because of a critique of this evidence that highlights the lack of direct association between fossil proxies (6). We would support a comprehensive study of starches present in the New Zealand flora, as has been conducted for other regions (7), and further exploration of other fossil proxies for taro and other economic plants (8), to improve taxonomic identification.

• 2) High-precision radiocarbon dates: The dates on peat reported from Motutangi, unlike the several dates on identified macrobotanical remains from Ahuahu, do not meet widely accepted criteria for high-precision dating, as they contain mixed carbon sources (9).

• 3) Description of crop ecosystems: Ancient crop ecosystems cannot be described without comprehensive analyses of biological remains from archaeological contexts. At Polynesian arrival, the Motutangi wetlands were dominated by the large-statured conifer Dacrydium cupressinum, requiring repeated firing to establish gardens, and Restionaceae (3), most likely Apodasmia similis, a rush which dominates the margins of regularly flooded estuaries or lakes and outcompetes other plants in nutrient-poor soils (10). This densely spreading rush likely posed difficulties for crop cultivation, although taro may have been competitive if grown in clumps over multiple seasons. The monocotyledon trees, Rhopalostylis and Cordyline, prevalent in the ICP fossil assemblages from Ahuahu, but absent from the Motutangi fossil records, indicate easily cleared forest with immediately workable, nutrient-rich soils. We also identify fossil pollen and seeds of several additional plants with economic value including leafy green vegetables (e.g., Rorippa divaricata and Sonchus kirkii), further indicating cultivation contexts. Furthermore, ditch irrigation, reticulation, and drainage features, similar to those described for Motutangi, were recently excavated on Ahuahu, at Waitetoke, with the fossil assemblages we present in our paper.

Finally, we agree the Little Ice Age may have induced changes in crop choices and cultivation strategies, perhaps linked to shifts in fire regimes and enhanced forest clearance (11), although direct evidence for such a link is lacking.