We investigated early spontaneous colonisation patterns during semi-dry grassland restoration at two sites in SE Austria. The sites were left to regenerate passively without addition of plant propagules on a former arable field and an apple orchard. The sites were prepared only by ploughing (arable field) or clear cutting (apple orchard) and subsequently mowed annually. We studied whether, four years after project initiation, target species from adjacent semi-dry grasslands had established at the restored sites. We asked: 1) Does passive restoration lead to the establishment of target species? 2) Do abiotic parameters or distance to reference sites explain early colonisation patterns? 3) Do plant traits predict the colonisation success of different species? At each site, we collected data in 4 m × 4 m plots, in which we sampled the vegetation, analysed abiotic parameters (soil potassium- and phosphorus-content, soil-pH, slope) and recorded the minimum distance to the reference site. We tested for correlations between abiotic variables, plant traits and colonisation success. Colonisation patterns were not driven by abiotic soil conditions but rather by nearest distance to the reference sites. In addition, the vegetation developed differently in the former arable field and the apple orchard. Competitive species of the Arrhenatherion and thermophilic ruderal associations dominated the early restoration stage at both sites. Passive restoration of semi-dry grasslands on former agricultural land is unlikely to succeed unless complemented by initial ploughing, nutrient stripping and addition of propagules of rare species. Preučevali smo vzorce zgodnje naselitve vrst med obnovo polsuhih travišč na dveh lokacijah v jugovzhodni Avstriji. Travišča smo prepustili pasivni obnovi brez vnosa propagul na nekdanjo obdelano površino in sadovnjak jablan. Obe površini smo predhodno obdelali tako, da smo polje preorali, sadovnjak pa posekali in nato kosili vsako leto. Po štirih letih od začetka projekta smo spremljali, ali se na obnovljenih površinah pojavljajo tarčne vrste s sosednjih polsuhih travišč. Zanimalo nas je: 1) Ali pasivna obnova omogoča naselitev tarčnih vrst? 2) Ali lahko z abiotskimi dejavniki ali oddaljenostjo ciljnega travišča razložimo vzorce zgodnje naselitve? 3) Ali lahko z rastlinskimi znaki napovemo uspešnost naselitve različnih vrst? Na vsaki lokaciji smo postavili poskusne ploskve 4 m × 4 m, kjer smo vzorčili vegetacijo, abiotske dejavnike (vsebost kalija in fosforja v tleh, pH tal, naklon) in izmerili minimalno oddaljenost od ciljnega travišča. Testirali smo korelacijo med abiotskimi dejavniki, rastlinskimi znaki in uspešnostjo naselitve vrst. Vzorci kolonizacije so bolj odvisni od bližine ciljnega travišča kot pa od abiotskih dejavnikov tal, poleg tega se je vegetacija razvijala drugače na nekdanji njivi kot v sadovnjaku. Na obeh rastiščih so v zgodnjih fazah obnove prevladovale konkurenčno uspešnejše vrste zveze Arrhenatherion in termofilnih ruderalnih združb. Pasivna obnova polsuhih travišč na nekdanjih obdelovanih površinah bo verjetno neuspešna brez predhodnega oranja, odstranjevanja hranil in dodajanja propagul redkih vrst